AMD | Bulldozer, Fusion, AM3+, FM1, and What's To Come

K? What happened to Black Edition? Getting way too close to Intel's naming scheme there. Ties right in with the SBe and IB naming.
Seems they're reserving it for the higher end desktop parts, like how the Athlons and Phenom II X4 840 didn't have that designation. It's just the choice of "K" that comes off as piggybacking on Intel's marketing.


Same Frequency! FX-8150 vs Phenom II X6, X4, and Intel i5 2500K
http://en.inpai.com.cn/doc/enshowcont.asp?id=7982&pageid=8095

Article Index

PAGE 01:-Introduce
PAGE 02:-Bulldozer architecture features brief review
PAGE 03:-The Test
PAGE 04:-Arithmetic Performance - Sandra 2011
PAGE 05:-Multimedia Performance - Sandra 2011
PAGE 06:-Encryption and decryption test - Sandra 2011
PAGE 07:-Multi-Core Efficiency Test - Sandra 2011
PAGE 08:-Memory bandwidth test - Sandra 2011
PAGE 09:-Operational Performance Testing - Wprime 2.05
PAGE 10:-Operational Performance Testing - Fritz Chess
PAGE 11:-File compression test - WinRAR 4.01
PAGE 12:-File compression test - 7-Zip 9.20
PAGE 13:professional rendering test - CINEBENCH R11.5
PAGE 14:-Video Encoding - H .264 Encoder V2
PAGE 15:-Benchmark Test- 3DMark Vantage
PAGE 16:-Game Performance - Crysis 2
PAGE 17:-Game Test- Battlefield: Bad company 2
PAGE 18:-Game Test- Colin McRae: DiRT 3
PAGE 19:-Summary
 
AMD FX-6100 - 3.3GHz base, 3.9GHz max. turbo
AMD FX-6200 - 3.8GHz base, 4.1GHz max. turbo

DH Special: the official details about the processor AMD's FX-6200
http://www.donanimhaber.com/islemci...-FX6200-islemcisi-hakkinda-resmi-detaylar.htm

Translation said:
AMD, Bulldozer-based FX processor family is preparing to offer the market the second 6-core model. Technical specifications of the new processor known as the FX-6200, price and release date upon the official details. Standard 2.4 GHz with the Turbo Core 2.0 technology in running and working up to 2.5 GHz Core frequency of the processor 6 MB level 2 that automatically and comes with memory size of 8 MB level 3. Compared with the previous model, the FX-6100 processor thermal design power of ascending to the base rate is notable, this is a significant increase of 95 Watt, depending on the update from 125 Watts ascended State.

As a multiplier unlocked so that the processor, AMD overclocking ability for applications according to the official performance of Intel's Core i5 performance, video-editing-2400 processor while leaving behind the FX-6100 processor manages to be at a rate of 10-15% faster. December 26, will be launched in the FX-6200, n will be the official sale price $ 175 abroad.
 
On Bulldozer in a Windows environment, with CMT operating similarly to SMT:

The Idea Behind AMD’s Bulldozer
http://www.tomshardware.com/reviews/fx-8150-zambezi-bulldozer-990fx,3043-3.html

In Intel Core i5 And Core i7: Intel’s Mainstream Magnum Opus, I brought up specific optimizations in Windows 7 that were the product of collaboration between Intel and Microsoft—notably, core parking. Windows 7 intelligently schedules to physical cores before utilizing logical (Hyper-Threaded) cores.

In theory, AMD could benefit from the same thing. If Windows were able to utilize an FX-8150’s four modules first, and then backfill each module’s second core, it’d maximize performance with up to four threads running concurrently. This isn’t the case, though. According to Arun Kishan, software design engineer at Microsoft, each module is currently detected as two cores that are scheduled equally. So, in a dual-threaded application, you might see one active module and three idle modules—great for optimizing power, but theoretically less ideal from a performance standpoint. This also plays havoc with AMD’s claim that, when only one thread is active, it has full access to shared resources. Adding just one additional thread could tie up those shared resources, even as multiple other modules sit idle.

Microsoft is looking to change that behavior moving forward, though. Arun says that the dual-core modules have performance characteristics more similar to SMT than physical cores, so the company is looking to detect and treat them the same as Hyper-Threading in the future. The implications there would be significant. Performance would unquestionably improve, while AMD’s efforts to spin down idle modules would be made less effective.


On Bulldozer doing one of the things many have known for years it was largely geared for:

AMD Designed Bulldozer For Large Virtual Machine Loads
http://www.informationweek.com/news/hardware/processors/231903448

AMD's new Bulldozer chip is a set of pragmatic tradeoffs that get a new processor to market before the Intel juggernaut can launch its next generation Ivy Bridge chip and lock up more of the server market. Whether Bulldozer can clear AMD's path back to prosperity remains to be seen, but it's a strong attempt that may find favor with those running heavily virtualized environments.

It's been criticized as underperforming on expected clock speed and invoking a previously failed approach to Pentium data pipeline design. And the way AMD achieved 16 cores by glomming two eight-core chips together, eliminating one of the floating point engines while keeping the two integer engines, will strike some as an inelegant approach. Nevertheless, it seems clear to me that AMD dispensed with some chip design barriers to go for the sweet spot in today's server market--maximizing a server's capabilities to serve as a virtual machine host.

To test that notion, I turned to Aaron Branham, director of infrastructure at BlueLock, a supplier of infrastructure as a service to startups and enterprises. His clients run lots of ESX Server virtual machines in his data center--BlueLock is one of a handful of VMware vCloud partners. His is a heavily virtualized environment; he routinely runs 100 virtual machines on his multi-tenant servers.
 
Several miles? Various sites have done direct comparisons between the Opterons, new and old, and the Xeons. Some complete with highly sensationalistic pro-, or anti-, titles and/or content, while others have been more sensible in their approach. Performance, price, consumption, future prospects, etc.

AnandTech's assessment was posted on the previous page.
Now it's a matter of seeing how things progress with Intel's X79 parts, AMD's revisions, and software optimizations.


AMD Opteron™ Processor-based Server Benchmarks
http://www.amd.com/us/products/server/benchmarks/Pages/benchmarks-filter.aspx

Technology Primer: AMD Opteron 6200 CPUs (Interlagos)
http://www.pugetsystems.com/labs/articles/Technology-Primer-AMD-Opteron-6200-CPUs-Interlagos-124



Review: AMD Interlagos - An Easy Upgrade?
http://www.xcpus.com/reviews/amd-interlagos-an-easy-upgrade

CPU test: AMD Opteron 6262 HE and 6276 with 16 core Bulldozer
http://www.tecchannel.de/server/pro...teron_6262_he_und_6276_mit_16_core_bulldozer/

SAP SD Standard Application Benchmark Results, Two-Tier Internet Configuration
http://www.sap.com/solutions/benchmark/sd2tier.epx

AMD Bulldozer 16-core server CPUs "trounce" Intel Xeon
http://blogs.computerworld.com/19273/amd_16_core_server_cpus_trounce_intel_xeon

AMD's Bulldozer server benchmarks are here, and they're a catastrophe
http://arstechnica.com/business/new...marks-are-here-and-theyre-a-catastrophe.ars/1

AMD smacks Xeon 5600s with Opteron 6200s - Xeon E5 counterattack imminent
http://www.channelregister.co.uk/2011/11/21/amd_opteron_versus_xeon_5600/

AMD set to Bulldoze the datacenter
http://www.zdnet.com/blog/datacenter/amd-set-to-bulldoze-the-datacenter/1111

Will 16-Core AMD Opteron Server Processors Keep AMD Relevant?
http://www.theinfoboom.com/articles/will-16-core-amd-opteron-server-processors-keep-amd-relevant/

AMD Opteron 6200 vs Opteron 6100 Value
http://blog.icc-usa.com/2011/11/28/amd-opteron-6200-vs-opteron-6100-value/

AMD Opteron 6128, 2 GHz (OS6128WKT8EGO) OEM / Unboxed Processor
http://www.epinions.com/review/AMD_...Cache_Tray_OS6128WKT8EGO/content_571615710852

AMD Opteron 6272 Benchmarks, Reviews, Open-Source Tests
http://openbenchmarking.org/s/AMD Opteron 6272



TOP500 List Releases: TOP500 List - November 2011 (1-100)
http://www.top500.org/list/2011/11/100

Cray Notches Another XE6-Cascades Super Deal
http://insidehpc.com/2011/12/01/cray-notches-another-xe6-cascades-super-deal/



Dell PowerEdge R815
http://www.zdnet.co.uk/reviews/sme-servers/2011/11/14/dell-poweredge-r815-40094404/



PR releases for reference:
New AMD Opteron(TM) Processors Deliver the Ultimate in Performance, Scalability and Efficiency
http://www.marketwatch.com/story/ne...nd-efficiency-2011-11-14?reflink=MW_news_stmp

New, Award-Winning AMD Opteron™ Processors Power World’s Elite Supercomputers
http://www.amd.com/us/press-releases/Pages/new-award-winning-amd-opteron-2011nov15.aspx
 
Arun says that the dual-core modules have performance characteristics more similar to SMT than physical cores, so the company is looking to detect and treat them the same as Hyper-Threading in the future.
AMD should have pushed for that from the start...

Eh, I thought Xeons still beat Bulldozers by several miles?
The opposite seem to be the case. German Heise:
2x Interlagos (Opteron 6276) vs 2x Westmere-EP (Xeon 5680)
DGEMM: 222-239 GFlops vs 144 GFlops
SPECInt_rate2006_base: 454 vs 349
SPECfp_rate2006_base: 337 vs 246
Linpack efficiency: 502 MFlops/Watt vs 311 MFlops/Watt
SPECPower-Benchmark: 1736 ssj_ops/Watt vs 1203 ssj_ops/Watt
 
More on the unlocked AMD A8-3870K and A6-3670K Fusion APUs that were posted earlier:

DH specific: AMD Fusion to A8 A6-3670K-3870K, release date and features, and pricing information
http://www.donanimhaber.com/islemci...zellikler-cikis-tarihi-ve-fiyat-bilgileri.htm

WEB TRANSLATION said:
Bulldozer architecture-based 6-core FX prepared to offer to the market on December 26-6200 processor, AMD Fusion processor family on the same date, and also goes by the multiplier unlocked for desktop systems K series Fusion processors will offer a taste of the industry.


The first details of the earlier haberlerimizde our K series new Fusion processors technical specifications and prices are found answers to questions about curiosity. Upon reaching is the fastest processor, based on information Fusion series quad-core will be the price tag of $ 135, A8-3870K. 3 GHz and 4 MB level 2 running in memory to the processor, which has kapasiteisne of dual-channel DDR3 1866MHz memory and comes with integrated graphics Radeon HD 6550D. the transaction volume and speed of 600 MHz parallel 400 running DirectX 11 compatible graphics unit at the same time with an external display support for Dual Graphics cards for use.


The second is unlocked multiplier AMD's Fusion processor A6-3670K also has similar to quad-core. 1.7 GHz and 4 MB level 2 running in memory to the processor that is capable of dual-channel DDR3 memory, similar to 1866MHz 443MHz up and running in parallel operation supports 320 x 6530D Radeon HD graphics with unit volume is still with the design of the processor's thermal power Fusion is going to be in a similar way 100 Watt A8-3870, and will be available from $ 115.
*click for 600px × 441px

 
AMD should have pushed for that from the start...
On the list of terrible decisions, and oversights made by AMD management. For their sake, they should hope the new leadership has better vision, and ability to execute, despite the relative lack of resources, market cap and blah-blah-blah.

On that note, a little birdie from Microsoft dropped this off:

An update to optimize the performance of AMD Bulldozer CPUs that are used by Windows 7-based or Windows Server 2008 R2-based computers is available
http://support.microsoft.com/kb/2592546/

Article ID: 2592546 - Last Review: December 14, 2011 - Revision: 1.0

This article introduces an update that optimizes the performance of AMD Bulldozer CPUs that are used by Windows 7-based or Windows Server 2008 R2-based computers. Currently, the performance of AMD Bulldozer CPUs is slower than expected. This behavior occurs because the threading logic in Windows 7 and in Windows Server 2008 R2 is not optimized to use the Simultaneous Multithreading (SMT) scheduling feature. This feature was introduced in the Bulldozer family of AMD CPUs.

Note This issue may occur when you use applications that run in multiple threads.
After you install this hotfix, Windows 7 and Windows Server 2008 R2 fully support AMD Bulldozer CPUs. Therefore, the CPUs operate at the expected performance level.
 
Interesting. Any chance that hotfix will have a meaningful impact?
It should. If you take a look at posts over the last few pages on Windows 8 benches, or different core configurations under Windows 7, you'll see improvements in most instances. People have reported anything from negligible improvements, to ~10%, or higher increases. This isn't a cure-all by any means, but along with improvements in the fab process, BIOS, software optimizations and so on, it's a step in the right direction.

When I have a moment I'll try to compile them into one post, and/or link to them in the OP.
 
There's no telling if very many tech sites will pick up on the Win 7 update, and re-run benches, since that's very time consuming. Perhaps when Windows 8 beta releases? Either way, until more professional sources can provide results of some sort, I'll post user accounts from various sites, just to get a picture of how things are shaping up.

Grains of salt at the ready:

cageymaru:

I played Saint's Row the Third for a few minutes with FRAPS running to check the fps. Driving laps around the nuclear power plant yesterday I got 60 fps as the max and dips as low as 32 fps. Today with the scheduler patch I got 60 fps as the max and 45 as the minimum. Seems to give the game a nice boost to stability.

1680x1050 and every option maxxed out.

mikezachlowe2004:





BF3 is a little different. You can see the top ss is before the patch. The bottom one is from after patch. There is a change in core utilization. I got a 15-20fps increase on battlefield 3 with my CF and eyefinity.went from 30-40 to 50-60. MW3 increased by 35-45fps for me too. Below shows MW3. First is before, second is after patch.





You can see here that MW3 uses it more as a 4 core.

I think what is going on here is that both cores in each module are being untilized to not share resources. One core in a module is using one instruction set while the other core in the same module is using a different instruction set. Kind of like the cores are woking without getting in each others way. So that two cores that are in the same module are not using the same resources at the same time. Thus, creating more throughput for each module. This is the best way I can describe what I think is going on. I can definitely feel more quickness on the desktop too. I will update when I find out more info.

*click for 6,088px × 1,080px

mikezachlowe2004:

With 3D advantage it seemed like fps were up but score doesn't really show that. CPU doesn't show much gain either. Now with BF3 and MW3 I could tell a difference. I tried to take screen shots of fps but it wouldn't work. [...]

I definitely notice increase in performance in gaming and I can see some quickness to the desktop as well. I will update as I find out more. I will get some more screen shots to upload when I mess around a little more later.

bmgjet:

Been doing some of my own testing.
Some bench marks are picking it up as 4core 8thread where others still see it as 8core 8thread.
Proformance test 7 still sees it as 8 core 8 thread and has gotten worse results with patch.
Integer and floating point are 0.3% worse where string sorting has seen a 5% increase and phyics a 1.8% increase.
Ram test saw a 2.1% increase but a 0.2% decrease in write speed.
Prime has seen a 0.133 improvement so now im into the 8s instead of low 9s.
Battle field 3 see it as 8 core still and has no change.
CS:S saw a 10fps increase [just] like if I run with 1 compute unit per core disabled in the bios.
Saints row 3 runs a lot more smoothly as well as Garys Mod when you have a lot going on.
Cine bench I got a 0.07 increase in in multi core and 0.02 increase in single core.
Serious sam 3 doesnt have slow downs in certain parts any more.
Going to be doing more testing with BF3 latter on to see if it still slows down when getting knifed.
Intel Burn in test Im now getting 51Gflop instead of 48
.
Been able to overclock some more as well another 90mhz on the same voltage where I only had 20mhz safety overhead before hand.

riska: (WEB TRANSLATION)

Till remains active, I cannot see some improvements with the ...

Before the fix



After the fix



*click for 1,920px × 1,080px

arthurs:

Well I can confirme improvements with this fix on FX6100@4000 in the x264 HD BENCHMARK 4.0 and luxmark 1.0

For x264 HD BENCHMARK 4.0

before fix

Results for x264.exe r1913
==========================

Pass 1
------
encoded 1442 frames, 114.26 fps, 3913.30 kb/s
encoded 1442 frames, 114.40 fps, 3913.30 kb/s
encoded 1442 frames, 113.70 fps, 3913.30 kb/s
encoded 1442 frames, 112.87 fps, 3913.30 kb/s

Pass 2
------
encoded 1442 frames, 28.00 fps, 3959.17 kb/s
encoded 1442 frames, 27.92 fps, 3958.75 kb/s
encoded 1442 frames, 28.02 fps, 3959.86 kb/s
encoded 1442 frames, 27.93 fps, 3960.67 kb/s


After fix

Results for x264.exe r1913
==========================

Pass 1
------
encoded 1442 frames, 119.43 fps, 3913.30 kb/s
encoded 1442 frames, 119.27 fps, 3913.30 kb/s
encoded 1442 frames, 118.96 fps, 3913.30 kb/s
encoded 1442 frames, 121.78 fps, 3913.30 kb/s

Pass 2
------
encoded 1442 frames, 31.42 fps, 3958.43 kb/s
encoded 1442 frames, 31.49 fps, 3958.93 kb/s
encoded 1442 frames, 31.53 fps, 3959.72 kb/s
encoded 1442 frames, 31.56 fps, 3959.45 kb/s

-------------------------------------------------

Luxmark 1.0 64bit Native CPU

Before
View attachment 44678

After

View attachment 44679

Luxmark 1.0 64bit OpenCL CPU

Before

View attachment 44682
After

View attachment 44683


Somehow it is working 5%-10% faster
Ö

*lack of attachments originate with original poster

alexmaia_br:

SAINTS ROW the 3rd:
Wow… here there is a BIG difference. The game plays much better after the patch. It’s more fluid, no stuttering
…

To take this picture, I ran around and found a spot where my fps would suffer more… somewhere easy to mark. I then proceed to wait for 5 minutes or so, taking pics whenever my fps was the lowest. Really, there’s nothing more to say here than… GREAT!
Even being aware there’s a problem with AMD drivers and low FPS, it’s perfectly playable now (it wasn’t before, at least for me)
. Here I saw the magic happening biggrin.gif

EDIT: low graphics card usage is more of a crossfire thing atm, I think. From steam forums:
But there is still hope since volition posted a new thread on the official SR3 PC forums stating[...]



BEFORE:


AFTER:


EDIT:
SERIOUS SAM 3:
Also solid improvements. Settings are everything on ultra, 1080p. I got the spot where I found my fdps took the greatest hit, just in the beginning of the game. Saved… then loaded and took pics without moving. The pics were taken during the period of 2-3 minutes, whenever it hit a lower grade I’d SC it. As with saints row, the game now uses my xfire config better, allowing for greater fps:

BEFORE:


AFTER:


*click pics for higher res

computerparts:

I wasn't expecting miraculous gains because I knew what this patch was meant to do. Seems to be doing it's job, no longer thread crapping across all cores.

Before


After


*click for higher res
 
The improvements are quite nice, it still draws too much power though. If they would be at Intel's 95W TDP levels I would consider one (if it would be available), but in this state I think i5 is still the better buy (and is way cheaper). Hell, even the i7 has a higher value IMO (prices for the 2600 are around 20-30 higher than for the 8150, not too much for a high-end system)
 
So Bulldozer was good all along?
Yes, and no. Many people overlook the areas where it excels, and focus solely on the ones where it needs work. Unfortunately for AMD, they aren't an easy "yes" recommendation across the board.

The patch wasn't fully tested, and since it was also supposed to be a 2-part patch, instead of the single patch that was released early, Microsoft pulled it.

http://support.microsoft.com/kb/2592546/

AMD and Microsoft are continually working to improve hardware and software for our shared customers. As part of our joint work to optimize the performance of “Bulldozer” architecture-based AMD processors we collaborating on a scheduler update to the Windows 7 code-base. The code associated with this KB is incomplete and should not be used.
Update: MS releases, yanks Win7 hotfix for Bulldozer scheduling
http://techreport.com/discussions.x/22179

Update II - 10:55AM: We've spoken with an industry source familiar with this situation, and it appears the release of this hotfix was either inadvertent, premature, or both. There is indeed a Bulldozer threading patch for Windows in the works, but it should come in two parts, not just one. The patch that was briefly released is only one portion of the total solution, and it may very well reduce performance if used on its own. We're hearing the full Windows update for Bulldozer performance optimization is scheduled for release in Q1 of 2012. For now, Bulldozer owners, the best thing to do is to sit tight and wait.
That makes sense as it should coincide with the release of the newer B3 stepping FX-8170, FX-8140, FX-6120, and FX-4120 CPUs (see below). So, the upcoming December/January B2 stepping FX-8120 (95-watt version), FX-8100 and FX-4170 reviews won't have completed patches, unless some sites decide to use the partial patch, anyway.



The improvements are quite nice, it still draws too much power though. If they would be at Intel's 95W TDP levels I would consider one (if it would be available), but in this state I think i5 is still the better buy (and is way cheaper). Hell, even the i7 has a higher value IMO (prices for the 2600 are around 20-30 higher than for the 8150, not too much for a high-end system)
While it won't be drastic, better scheduling, and software level core management, will have a positive impact on performance-per-watt. That's to go along with the expected B3 stepping improvements due in Q1, ahead of the revised "Bulldozer Enhanced," aka Piledriver, revisions in Q3/Q4.

Also, remember the actual MSRP without the retail markups: http://www.amd.com/us/press-releases/Pages/unlock-your-record-setting-2011oct12.aspx
  • FX-8150: $245 suggested retail price (U.S.)
  • FX-8120: $205 suggested retail price (U.S.)
  • FX-6100: $165 suggested retail price (U.S.)
  • FX-4100: $115 suggested retail price (U.S.)
You can argue those are still not competitive, but markups are getting factored in as the standard prices by many. The FX-8150 goes for $240 on Amazon, and the FX-8120 sells for ~$200. There's very little difference between the two. The FX-4 and FX-6 also go for sub-MSRP. Users have had the 95w version of the FX-8120 for several weeks now, and it should be available any week now.


Proposed Bulldozer B3 stepping release schedule:

 
125w FX-8120 power draw relative to 95w Core i5 2500 (K, and non-K), 95w Core i5 2600K, 125w Phenom II X4/X6, and 125w FX-8150:


Power Consumption (lower disparity)
http://www.xbitlabs.com/articles/cpu/display/amd-fx-8120-6100-4100_8.html
Power Consumption (higher disparity)
http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/9
Fluctuating FX-81xx power use figures across multiple sites' reviews largely attributed to bad/immature Global Foundries 32nm process, more so than disparity in each sites' specific testing methods, or choice of components.



95w FX-8120 power draw relative to 125w Phenom II X6 1100T

Power Consumption Test
http://forums.ninjalane.com/viewtopic.php?f=3&t=3036
http://www.wretch.cc/blog/windwithme/985727
*CNQ = AMD Cool'n'Quiet™ Technology (like Intel's power save features)

1100T Enter to OS Desktop (Enable CNQ) - 81W ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, FX-8120 Enter to OS Desktop (Enable CNQ) - 74W


1100T Enter to OS Desktop (Disable CNQ) - 103W ,,,,,,,,,,,,,,,,,,,,,,,,,,,, FX-8120 Enter to OS Desktop (Disable CNQ)- 93W


1100T Full Speed (LinX 0.6.4) - 238W ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, FX-8120 Full Speed (LinX 0.6.4) - 166W ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, FX-8120 Full Speed (LinX 0.6.4) - 157W (second test)


For reference: 125w FX-8150 power consumption relative to 130w Core i7 990x, and 130w Core i7 3960x. All three CPUs at 32nm. All three CPUs having greater than 4 standard cores, by any definition.

http://www.xbitlabs.com/articles/cpu/display/core-i7-3960x-3930k_11.html#sect0
http://www.pcper.com/reviews/Proces...Quad-Core-Sandy-Bridge-E-under-300/Power-Cons
http://www.hardwarecanucks.com/foru...ndy-bridge-e-core-i7-3960x-cpu-review-21.html
http://www.anandtech.com/show/5091/...-bridge-e-review-keeping-the-high-end-alive/7
http://www.technic3d.com/review/cpu-s/1361-intel-core-i7-3930k-und-i7-3820-im-kurztest/7.htm

Stock, and overclocked power consumption *note: below FX-8120 is 125w version*

http://www.hardocp.com/article/2011/11/14/intel_core_i73960x_sandy_bridge_e_processor_review/8
http://www.xbitlabs.com/articles/cpu/display/core-i7-3820_8.html#sect0
95w Intel vs 130w Intel

Isolated CPU power consumption

*note: 95w FX-4100, and 95w FX-6100 compared to 95w Intel CPUs.
Going by the idle, and load differences between the FX-4100, and FX-6100, one can extrapolate a rough idea of the power consumption of an average 95w FX-8100/FX-8120/FX-8140 and so on.

http://ht4u.net/reviews/2011/intel_sandy_bridge_e_hexa_core/index17.php
 
I thought I posted this earlier, but here's a follow up to some of the earlier discussion about how core configs, proper thread scheduling, and erratic core behavior can effect performance, and power draw.

*everything below is a direct quote, with no additional commentary


Analyzing Core Parking in Windows 7 and Windows 8 in relation to AMD Bulldozer performance (in English)
http://m.diit.cz/clanek/analyzing-c...erence-in-windows-7-and-windows-8/39100/37378

Article Chapters:

1. How to enable the Core Parking feature to be modified in user-friendly way
2. WinRAR vs. Core Parking vs. Core i7 in Windows 7 and Windows 8
3. Is Bulldozer module dual-core, or single-core with Hyper-Threading on steroids?
4. Core Parking difference in Windows 7 and Windows 8
5. Conclusion
Core Parking difference in Windows 7 and Windows 8
There is another difference in Core Parking between Windows 7 and Windows 8. Look at default Core Parking for Core i7 (I'm sure you'll understand that "zaparkováno" in Czech means "parked" in English ;-).

Difference between Core Parking in idle: Windows 7 (left) and Windows 8 (right)

The difference is that Windows 7 parks all "hyper-threads" while Windows 8 parks all cores (including HT) excluding the first one. Windows 8 is also able to park AMD FX processor modules the same way:

Core Parking of AMD Bulldozer 8-core processor: Windows 7 (left, unable to park anything), Windows 8 (right)

Since the Core Parking feature is generally for energy saving and not for performance improvement, the AMD FX processor is normally favored in reviews by the Windows 7 operating system. In spite of that fact AMD Bulldozer failed to show some great performance, with the one exception, which is WinRAR.

We would like to show you what happens when you would like to change the affinity of cores to particular process in Windows 7 and 8 when Core Parking is enabled. As you will see, it makes an extraordinary mess in the performance and cores/threads utilizing. Not working really the way you would expect ;-). (The graph shows the encoding process, orange line represents 2nd pass while the beginning shows 1st pass of encoding).

Core Parking ON, affinity 11110000, Windows 7 (left), Windows 8 (right)

You would probably guess that Windows 8 works as expected in this situation while Windows 7 does something odd, namely park the threads that are assigned to the application. But look what happens in the 10101010 situation (ignore the spikes on the right picture, Windows 8 is far from final and far from perfect):

Core Parking ON, affinity 10101010, Windows 7 (left), Windows 8 (right)

Roles changed, Windows 7 works as expected while Windows 8 parks some threads that are set to work. Of course it has a big influence of the results:

Needless to say that very few people will probably set the affinity of process like that, we did it just for the demonstration purposes. And if someone would really want to do that, he will probably turn off the Core Parking at all.
 
NP, Haz. Always good to know some people find them useful.


So, they did use the "Black Edition" branding, after all.


AMD A8-3870K & A6-3670K Reviews

bit-tech
http://www.bit-tech.net/hardware/cpus/2012/02/06/amd-a8-3870-review/1
Bjorn3D
http://www.bjorn3d.com/articles/AMD_A8-3870K_Black_Edition_Llano/2164.html
CowcotLand
http://www.cowcotland.com/articles/993/test-processeur-amd-a8-3870-k.html
Guru3D
http://www.guru3d.com/article/amd-a8-3870k-review/
HardwareCanucks
http://www.hardwarecanucks.com/foru...2-amd-a8-3870k-unlocked-llano-apu-review.html
Hi Tech Legion
http://www.hitechlegion.com/reviews/processors/16501-amd-a8-3870k-llano-apu-review
HotHardware
http://hothardware.com/Reviews/AMD-A83870K-Unlocked-Llano-QuadCore-APU-/
HT4U
http://ht4u.net/reviews/2011/amd_a8_3870k_llano_apu/
Inpai
http://www.inpai.com.cn/doc/hard/164535.htm
TheLab.Gr
http://www.thelab.gr/vgas-motherboards-cpus-ram-reviews/amd-a8-3870k-review-103642.html (Greek, 13-page)
lab501.ro
http://lab501.ro/procesoare-chipseturi/amd-a8-3870k-lets-overclock
Lostcircuits
http://www.lostcircuits.com/mambo//index.php?option=com_content&task=view&id=104&Itemid=1
Madshrimps
http://www.madshrimps.be/articles/article/1000288/
ocaholic
http://www.ocaholic.ch/xoops/html/modules/smartsection/item.php?itemid=669&sel_lang=english
PC Watch
http://pc.watch.impress.co.jp/docs/topic/feature/20111220_500296.html
[Phoronix]
http://www.phoronix.com/vr.php?view=16881
http://www.phoronix.com/vr.php?view=16880
Planet 3DNow!
http://www.planet3dnow.de/vbulletin/showthread.php?t=401434 (German, 7-page)
PureOC
http://www.pureoverclock.com/article1427.html
Red and Blackness Mods
http://rbmods.com/amd-a8-3870-fm1-cpu/
TechwareLabs
http://www.techwarelabs.com/amd-llano-a8-3870k-apu-review/
TweakPC
http://www.tweakpc.de/hardware/tests/cpu/amd_a8-3870k_llano_apu/s01.php
Vortez
http://www.vortez.net/articles_pages/amd_a8_3870k_review,1.html


AMD A8 3870K vs. Intel Core i3 21xx

Hardware Secrets
http://www.hardwaresecrets.com/article/A8-3870K-vs-Core-i3-2105-CPU-Review/1470
T-Break
http://tbreak.com/tech/2012/02/amd-a8-3870k-black-edition-review/
X-bit Labs
http://www.xbitlabs.com/articles/cpu/display/amd-a8-3870k.html


*click for 1,795px × 927px
 


HD 7970: Bulldozer vs. Sandy Bridge vs. Nehalem
http://www.techpowerup.com/reviews/AMD/HD_7970_CPU_Scaling/

In this review we will pit three CPUs from the popular ~$250 segment against each other using the Radeon HD 7970:

  • AMD FX-8150: This is the company's current flagship Bulldozer processor. It features four execution units that show up as eight cores in Windows, its Turbo clocks up to 4.2 GHz and it has plenty of cache, the processor is priced competitively at $270.
  • Intel Core i5-2500K: The processor most gamers and overclockers use. Thanks to its "K" suffix it enables significant overclocking unlike the cheaper Sandy Bridge processors, yet comes at an affordable price of $220. Actually this makes it the cheapest processor in today's test group.
  • Intel Core i7-920: While it does not use a sexy 32 nm production process and is based on Intel's last generation Nehalem architecture, it still packs a punch especially when overclocked. It is also an affordable choice if you are looking to run a multi-GPU gaming rig at full 2x PCIe x16 without breaking the bank (LGA 2011).
01 - Introduction & Specifications
02 - Test Setup
03 - Aliens vs. Predator
04 - Batman: Arkham City
05 - Battlefield 3
06 - BattleForge
07 - Civilization 5
08 - Crysis 2
09 - DiRT 3
10 - Dragon Age II
11 - Hard Reset
12 - Metro 2033
13 - STALKER: Call of Pripyat
14 - StarCraft II
15 - Total War: Shogun 2
16 - The Elder Scrolls: Skyrim
17 - World of Warcraft: Cataclysm
18 - 3DMark 11
19 - Unigine Heaven 2.0
20 - Performance Summary
21 - Value and Conclusion





Value and Conclusion
It looks like the clear winner of our little skirmish is Intel's "Sandy Bridge" Core i5-2500K. The processor has a significant performance advantage over our other contenders thanks to the combination between an efficient microarchitecture, high clock speed and Turbo Boost. At our lowest resolution of 1024x768 we see about 20% improved framerates when compared to the i7-920 and the AMD FX-8150. However, as resolution increases, and our tests become more GPU limited than CPU limited, the performance lead shrinks to below 10%, which is still noteworthy.

When looking at AMD's Bulldozer at low resolutions, which is where each CPU can prove its worth, we see the FX-8150 showing more weakness than strength, despite its eight CPU cores. That does not mean that Bulldozer is hopeless for gaming. Once resolution and AA is cranked up, even the fast HD 7970 GPU becomes the limiting factor in most games and we see very decent performance. Unfortunately platform cost of the Bulldozer setup is similar or higher when compared to the Intel 2500K which does offer better performance overall. In its current state Bulldozer is a very tough sale to any member of the gaming crowd. I could imagine it being interesting for users who want to use other benefits of the platform, like multi GPU gaming with two graphics cards connected via PCI-Express x16 each - Sandy Bridge LGA 1155 supports only 2x x8. Another incentive for Bulldozer could be the native six SATA 6 Gbps ports with RAID 0, 1, 5, 6 support, which would make a hell of a server. Nevertheless, the average gamer is better off with Intel's LGA 1155 Sandy Bridge platform.

Our aging Core i7-920 is still doing respectably well. So if if you still have a first-generation Core i7, there's no need for you to upgrade.
 
http://forum.beyond3d.com/showpost.php?p=1610388&postcount=166

Scaled comparison of Llano and Trinity, using the I/O pads on the left side for reference:



Some observations on the layout of the SIMD multi-processors -- the placement of the register file banks in the ALU array is different in Trinity, as well as the whole layout of the texture unit.

Here are the differences (so far) on the CPU side -- BD vs. Piledriver cores:



Those banks are most probably the pre-decode bits (used for the BTB, branch selector, end bits & etc.), that AMD has been using ever since the first K7 architecture to aid the instruction decode flow. And since these are located in the branch prediction area of the front-end block, I guess AMD is aiming at improving namely this aspect of the architecture.
For more on BTB and brand prediction, The Real World Technologies "AMD's Bulldozer Microarchitecture" article, found in the OP, covers it on page four.

http://www.realworldtech.com/page.cfm?ArticleID=RWT082610181333&p=4



edit:
http://forum.beyond3d.com/showpost.php?p=1610608&postcount=177

This is comparison of the IGP "uncore" sections of Llano and Trinity -- SIMDs are cut out too. Trinity's section takes 40% more area, compared to Llano's.

 



AMD Demos Next Generation Trinity Fusion APU, CES 2012
http://www.youtube.com/watch?v=lsmTDb-Mlws


AMD Fusion APU and Graphics Update: Trinity, 7000M Mobile GPUs, Android on X86 and Lightning Bolt
http://hothardware.com/News/AMD-Fus...00M-Mobile-GPUs-and-Lightning-Bolt-In-Action/

First up we have a video showing an upcoming Trinity APU in action. We don’t want to give away the “ah ha” moment in the video, but what’s interesting—along with the actual demo—is that AMD already has Trinity up and running, and running well as you’ll see. Particulars like clock speeds and the GPU configuration weren’t disclosed, but we can tell you that the Trinity APU in the notebook used during the demo was a 17w variant, and as AMD has already disclosed, the Trinity APU sports quad Piledriver cores, an update to Bulldozer that should offer better performance though not only architectural enhancements but frequency increases as well.
In a backroom session in which we weren’t able to snap any photos, AMD also showed us some interesting I/O technology, currently codenamed “Lightning Bolt” (not to be confused with Intel’s Thunderbolt). With Lightning Bolt, AMD aims to offer the ultimate docking station of sorts for AMD-based mobile devices. Using a $40-ish hub, users will be able to connect a single DisplayPort 1.2 cable to a mobile PC to not only charge the system, but to connect it to up to four external displays and multiple external USB 3.0 devices.


AMD Demos Android 4.0 on Fusion APU Tablet
http://www.youtube.com/watch?v=5hwbtmIvE28

Another very cool demo in the AMD booth featured an MSI-built, AMD-based tablet PC running Android 4, Ice Cream Sandwich. The 10-inch MSI WindPad used in the demo was outfitted with an AMD Z-Series Z-01 (1.0GHz) CPU, 4GB of DDR3 Memory, and Radeon HD 6250-series graphics. As you’ll see in the video, ICS ran well on the tablet and was smooth and responsive. Although still an early alpha that wasn’t fully functional, hardware acceleration was working, as is evidenced by the silky smooth graphics and scrolling.






AMD Shows Off Trinity APU Die And Trinity Powered Notebook
http://www.pcper.com/news/Cases-and-Cooling/AMD-Shows-Trinity-APU-Die-And-Trinity-Powered-Notebook

A die shot of the upcoming Trinity APU (Accelerated Processing Unit) sitting next to a USB flash drive. Specifications of the Trinity APUs have not yet been released by AMD; however, if this leak holds true the Trinity APUs will have either two or four Piledriver CPU cores and TDP (thermal design power) of 65 W, 100 W, and 125 Watts (depending on particular chip). Clock speeds will further vary between 2.2 and 3.8 GHz at stock speeds (will run a bit faster with Turbo Core 3.0). The GPU aspect will be clocked between 563 MHz and 711 MHz and is based on the VLIW4 technology of the Cayman graphics Cards (69xx). They estimate that it will deliver up to 30% more performance versus current Llano chips and will support all the fancy new X86 instruction sets like AVX and AES-NI. A nice boost and hopefully the real specifications will come close to this (or be even better, of course).


Update: Another interesting bit of information is that AMD will have a low power Trinity APU with a TDP of 17 watts and will supposedly deliver the same level of performance as the current Llano chips (that draw twice the power).

Update: AMD has stated Trinity will deliver a 25% increase in CPU performance and a 50% increase in GPU performance. Stay tuned to PC Perspective for more Trinity info as it develops.
 
Software optimization guide for Trinity/Piledriver/Bulldozer Enhanced.

Software Optimization Guide for AMD Family 15h Processors
http://support.amd.com/us/Processor_TechDocs/47414_15h_sw_opt_guide.pdf


Dozer - AMD Fam 15h - general info thread
http://www.planet3dnow.de/vbulletin/showthread.php?t=317639

WEB TRANSLATION said:
http://www.planet3dnow.de/vbulletin/showpost.php?p=4550397&postcount=410



There stands resulting HyperTransport I read that you will continue ServerCPUs (Family 15h Model 20-2Fh) are connected with each other via HyperTransport .... but if necessary via PCIe with the "chipset" and were not too far off the 10Nodes/20Cores (rather moved) Funny thing is also that to be eligible for Family 15h 2Fh Model 20-A Quad-Channel DDR3, which may also be a great for THE could indicate that no more base Gxx 2 THE per carrier would otherwise be yes because it is an 8-channel memory interface for the socket Gxx ...

 
CES: TDP of 17W for the APU Trinity!
http://www.hardware.fr/news/12078/ces-tdp-17w-l-39-apu-trinity.html

WEB TRANSLATION said:
AMD told us that three formats were planned, adding details about the level of performance is expected to approximate that Trintiy from Llano:

- Desktop: 65W and 100W, 15% at 30% CPU and GPU at compared to Llano desktop
- Mobile: 35W, 25% level CPU and GPU + 50% level compared to mobile Llano (35 and 45W)
- Ultra-(BGA): 17W, performance similar to mobile Llano 35W
WEB TRANSLATION said:
AMD insists to have worked in-depth energy efficiency of Trinity, so that it can significantly increase performance when the thermal envelope must be compressed as much as possible, as is the case with laptops. AMD may also benefit from the experience of facing the manufacturing process 32nm SOI GlobalFoundries and improving it over time.

The ultra-mobile version of the 17W TDP is that we expect the most, as it was sorely lacking in AMD Llano. An APU that will allow AMD to fight against Ultrabook Intel, which unfortunately will not be a level playing field given the offensive about it in terms of communication and the fact that s' acts of a brand that Intel has well protected.

AMD responds to this issue that consumers do not care a mark as qu'Ultrabook, but the experience offered by the products and their design. If AMD does not totally wrong, this does not change the fact that machine-based Trinity, which correspond to Ultrabook based Ivy Bridge will face numerous trade barriers, Intel teaming up with more retailers to offer specific Ultrabook rays. It also remains to be seen whether laptop manufacturers can be convinced to produce such high-end designs based Trinity. Breaks are announced on certain types of aluminum chassis for example, and some production lines have been funded by Intel ...

Nevertheless, AMD seems to have Trinity with a promising APU in the mobile world.




Features of AMD Piledriver processors
http://www.cpu-world.com/news_2012/2012011101_Features_of_AMD_Piledriver_processors.html

A few days ago AMD posted new technical document, titled Software Optimization Guide for AMD Family 15h Processors (PDF file). Although the guide is intended for developers, it includes some useful information about AMD Family 15h, i.e. Bulldozer. The document not only describes features of current Bulldozer processors, which have model numbers 00h - 0fh (0xh), but also refers to two future generations with model numbers 10h - 1fh (1xh), and 20h - 2fh (2xh). We believe that microprocessors with 10h and higher model numbers will be based on Piledriver cores.

The optimization guide is quite large. References to different features are scattered across the document, so we did our best to find all relevant information. There will be some features, applicable to both 1xh and 2xh CPUs, such as support for 16-bit floating point numbers, and addition of VCVTPH2PS and VCVTPS2PH instructions, used to convert to and from new 16-bit floating-point type. The processors will also incorporate FMA3, Bit Manipulation Instructions (BMI) and Trailing Bit Manipulation instructions, or TBM. In addition to this, there will be some other improvements, such as increased depth of FP load queue, and larger size of level 1 data TLB. Also, latencies of some instructions were reduced.

Microprocessors with model 1xh will have up to 2 modules, or 4 cores, and will lack L3 cache. These characteristics match upcoming Trinity core, that will be utilized by mobile AMD chips. Model 1xh CPUs will also have enhanced IOMMU, or IOMMU v2, that will improve access of I/O devices to system memory, and will add such features as direct access to user I/O space, and interrupt remapping and filtering.

Processors with model number 2xh will have up to 5 modules, or up 10 cores, and support quad-channel DDR3 memory. Like Bulldozer CPUs, which are aimed at both desktop and server markets, 2xh may also target both markets. We suspect that 2xh parts will be released as "Vishera" CPUs on desktop, and "Terramar" and "Sepang" processors for servers. It is possible that server Terramar Opterons, that will integrate two dies on a chip and have up to 20 cores, will support 8 memory channels.

The optimization guide also mentions model 30h - 3fh, and 40h - 4fh processors, however it doesn't contain any details on these chips. The summary of all features, referenced by the optimization guide, is provided below:
 
Apparently, here's the full Windows 7 Bulldozer patch (although, it still seems as if one of the patches may not be not fully complete?).

It's 1 part scheduler update, 1 part core parking addresser.


Early Results Achieved with AMD FX Processor Using Windows® 7 Scheduler Update
http://blogs.amd.com/play/2012/01/1...-processor-using-windows®-7-scheduler-update/

January 11th, 2012

Some of you may remember that AMD FX processors use a unique dual-core module architecture codenamed “Bulldozer”, which current versions of Windows® 7 were not specifically architected to utilize. In essence, for those with an AMD FX-8150 Processor, for example, Windows 7 sees the eight available cores and randomly assigns threads to them.

In initial testing of the upcoming Windows 8 operating system, we’ve seen performance improvements of up to 10% in some applications, when compared to Windows 7. This is because the system correctly recognizes the AMD FX processor architecture and cores. Thanks to close collaboration between Microsoft and AMD, Microsoft recently completed back-porting some of the Windows 8 scheduler code for AMD FX processors into a hotfix for Windows 7.

So if you have an AMD FX processor, here’s what you can do to update your version of Windows 7:

1) Download the scheduler update (KB2645594) and install. This will tell the scheduler that your AMD FX processor contains dual-core modules (in fact this is similar to the SMT path that the other guys use). In essence, threads 1-4 now get assigned to their own module first.

2) Download the core parking scheduler update (KB2646060) and install. This will prevent Windows 7 shutting down unused cores prematurely when there are threads to be assigned (there’s a performance penalty parking and then un-parking a core).

The best possible cases for improvement are applications that use ½ cores in your AMD FX processor. In our testing using the AMD FX-8150 processor, we found the best improvement in wPrime, Left 4 Dead 2, and Lost Planet. Below you’ll find links to the patches:

http://support.microsoft.com/kb/2646060

http://support.microsoft.com/kb/2645594

Our testing shows that not every application realizes a performance boost. In fact, heavily threaded apps (those designed to use all 8 cores), get little or no uplift from this hotfix – they are already maxing out the processor. In other cases, the uplift averages out to a 1-2 percent uplift. But heck, it is free performance, and this is the scheduler model that will be used in Windows 8 (along with some further enhancements), so why not add it to your list of downloads?

If you have an AMD FX processor, head over to Microsoft and grab the free updates.


An update that selectively disables the Core Parking feature in Windows 7 or in Windows Server 2008 R2 is available
http://support.microsoft.com/kb/2646060

This article introduces an update that enables the hotfix installer to selectively disable the Core Parking feature in Windows 7 or in Windows Server 2008 R2 if an AMD FX, AMD Opteron 4200 Series, or AMD Opteron 6200 Series processor is installed.

Currently, the CPU Power Policies that are used by Windows 7 and Windows Server 2008 R2 are not optimized for the dual core AMD Bulldozer module architecture. This architecture is found on AMD FX series, AMD Opteron 4200 Series, and AMD Opteron 6200 Series processors. This can result in decreased system performance with multithreaded workloads in lightly-threaded environments.

When this update is installed, Bulldozer modules will be less likely to achieve the C6 power state. This potentially results in increased power consumption in more lightly-threaded environments.

Important If you apply this update, you cannot revert the settings by uninstalling this update. This update should only be installed on computers that have KB2645594 installed.
An update is available for computers that have an AMD FX, an AMD Opteron 4200, or an AMD Opteron 6200 series processor installed and that are running Windows 7 or Windows Server 2008 R2
http://support.microsoft.com/kb/2645594

This article introduces an update for computers that are running Windows 7 or Windows Server 2008 R2 and that have a processor from one of the following series installed:

  • AMD FX
  • AMD Opteron 4200
  • AMD Opteron 6200

Currently, the CPU scheduling techniques that are used by Windows 7 and Windows Server 2008 R2 are not optimized for the AMD Bulldozer module architecture. This architecture is found on AMD FX series, AMD Opteron 4200 Series, and AMD Opteron 6200 Series processors. Therefore, multithreaded workloads may not be optimally distributed on computers that have one of these processors installed in a lightly-threaded environment. This may result in decreased system performance for some applications.

Important The power management policy for processor core parking on your computer may prevent this update from producing the desired performance increase. To disable core parking, install the hotfix that is described in Microsoft Knowledge Base (KB) article 2646060.
edit
Direct links:

KB2645594 - 32-bit
http://hotfixv4.microsoft.com/Windo.../Fix387870/7600/free/441064_intl_i386_zip.exe

KB2645594 - 64-bit
http://hotfixv4.microsoft.com/Windo...2/Fix387870/7600/free/441065_intl_x64_zip.exe


KB2646060 - 32-bit
http://hotfixv4.microsoft.com/Windo.../Fix391262/7600/free/441279_intl_i386_zip.exe

KB2646060 - 64-bit
http://hotfixv4.microsoft.com/Windo...2/Fix391262/7600/free/441280_intl_x64_zip.exe


 
Every bit helps. I'm still waiting on some (hopefully thorough) professional reviews with the complete, 2-part Win 7 patch. Until then, I might throw up some more user accounts.


AMD's "Trinity" APU comes to the "middle of the year" (update)
http://www.computerbase.de/news/2012-01/amds-trinity-apu-kommt-zur-mitte-des-jahres/

WEB TRANSLATION said:
Update

At the CES show presentation in the notes the objectives that AMD wants to achieve with "Trinity". The 25/50 percent more power you want to bid on a 35-watt notebook platform than its predecessor. These are predictions, based on PC Mark Vantage benchmark for productivity and part of the 3DMark Vantage CPU to the graphics unit.




AMD’s Answer to Ultrabooks: Make ‘Em Cheap
http://techland.time.com/2012/01/11/amds-answer-to-ultrabooks-make-em-cheaper/

At AMD’s meeting room, Leslie Sobon, the company’s vice president of global product and outbound marketing, showed me a prototype by Compal–the Taiwanese company that manufactures laptops for many major brands–but wouldn’t let me take pictures of it. The laptop was a respectable 0.71 inches thick. And although it was definitely heavier than many of the Ultrabooks I’ve played with at the show, the kicker is the price: AMD is targeting $500 for laptops like this once device makers start producing them.

That’s a huge difference compared to the $1,000-and-up Ultrabooks that have dominated the show floor at CES. Samsung’s gorgeous Series 9 Ultrabook, for instance, will be priced at $1,399 and up. HP’s Spectre, also stunning with its glass-clad lid and palm rest, is $1,399. There’s definitely an opening for cheaper Ultrabook alternatives.

AMD already has a laptop that comes close to the prototype I saw. It’s an Asus notebook (pictured) selling for $450 when it arrives in the United States later this year. But it’s thicker and heavier than the next wave of laptops AMD plans to power, and it’s not as capable. The next generation, which should arrive around mid-year, will be able to handle multi-monitor output, even with power-intensive tasks such as gaming and video happening at the same time.
 
HardWare.fr ran a few tests on December 19, with the old 1-part patch that was released early.

They've yet to post detailed results of the new 2-part patch.


AMD FX-8150 and patch: what performance?
http://www.hardware.fr/focus/57/amd-fx-8150-patch-quelles-performances.html

19/12/2011

1920 * 1080

WEB TRANSLATION said:
The most important gain is obtained under Crysis 2 with 9.5% best, followed by 7.1% on the first pass of x 264 and more (which gives a 3.5% gain in total) and 5.1% under MinGW. There is however a decrease of 5% in WinRAR, while in many cases performance will not significantly move. On application average gain of 0.5%, and on the average of the 3D games leads to a gain of 3.1%.
WEB TRANSLATION said:
You understand it, this patch Windows transfigure not AMD FX. Performance improvement is however significant in some cases and can hope that the final output of the two patches scheduled for the first quarter of 2012 will win even a few percent.
*5% drop in WinRAR performance due to patch #2 not being available at time of early testing. See below 2-part patch comparison, for results in WinRAR, Fritz, and others.





Free performance boost, Microsoft bulldozer patch test
http://www.expreview.com/18212-2.html

2012-01-13

WEB TRANSLATION said:
Here we assume that changes value over 1.5% is counted as a promotion or performance degradation.
Left Column= Single patch (thread scheduling)
Right Column = Both patches (thread scheduling, and core parking)

Black = Minor gains, or losses
Red = More pronounced gains
Grey = More pronounced losses

Single patch:
WEB TRANSLATION said:
Gray highlighted in part three is the performance degradation is serious, especially WinRAR and PDNbench image tests, 13-30% the difference can be explained by the non-error, install, uninstall, KB2645594 patch times repeatedly, the test result is the same order of magnitude, a difference is still very large.

Judging from the overall test result, KB2645594 patch performance hasn't changed, but in the 2-4 thread tests, Fritz, Blender, CINEBench, and wPrime does more than 2% enhancement, can reach a maximum of 7%, optimizing allocation of threads can improve performance.
Both patches:
WEB TRANSLATION said:
Better than the first test the performance of all tests even if there is no obvious performance boost, but not dropped, wPrime, Fritz, Blender, CINEBench 2-4 thread tests still slightly increased range between 2-7.5% and raise an average of 4.3%.

Judging from test results, separate installation of the first patches are up and down, and two patch installed is better, even if not improve did not fall, so it is recommended you install both patches, patch KB2645594 patch first and then call KB2646060 on.

Summary:

Bulldozers long raucous performance patch is finally coming, final test and the results prove that two patch did improve the performance, of course, improving small proportion, averaging only about 4.3%, and mainly in use 2-4 Threading test in the overall performance has not changed.

edit: English version

Test for AMD Bulldozer Hotfixes from Microsoft
http://en.expreview.com/2012/01/19/test-for-amd-bulldozer-hotfixes-from-microsoft/20477.html
 
I don't know if this is the right place to ask, but you chaps seem knowledgeable. I'm about to switch my mobo out for an AM3+ socket mobo and I'll be buying a new processor. I'm looking to spend £200-£250 on the processor, is there an obvious favourite I should be looking for? Anything I should avoid like the plague? I only really use my PC for gaming.

Any help would be appreciated.
 
So shit still sucks? I hate when I'm right about stuff like this. I wanted BD to be the bees knees (wtf does that mean?!)
It's still a pretty significant jump in performance. You don't get 4-7% CPU performance jumps through a software update every day.

What I'm much more interested in is seeing the processor rematched against Intel's offerings to see where they really stack up now.
 
I don't know if this is the right place to ask, but you chaps seem knowledgeable. I'm about to switch my mobo out for an AM3+ socket mobo and I'll be buying a new processor. I'm looking to spend £200-£250 on the processor, is there an obvious favourite I should be looking for? Anything I should avoid like the plague? I only really use my PC for gaming.

Any help would be appreciated.
If you're spending that much on a processor, you should grab an Intel 2500k TBH

It's still a pretty significant jump in performance. You don't get 4-7% CPU performance jumps through a software update every day.

What I'm much more interested in is seeing the processor rematched against Intel's offerings to see where they really stack up now.
True
 
I read an article about a month ago about how the CPU doesn't matter for current games when playing 1080p+, none of the quad cores (doesn't matter if AMD or Intel) was miles ahead. So I would go for the Phenom II X4 960T BE for about 100€ or the Phenom II X6 1055T 95W for about 140€. It still isn't clear if Piledriver cores will come on AM3+, so I wouldn't waste too much money on the board either.

If you're spending that much on a processor, you should grab an Intel 2500k TBH
Or this or wait 3 months for Ivy Bridge. But as I wrote, I would only consider these if you want to squeeze out every fps possible.
 
I read an article about a month ago about how the CPU doesn't matter for current games when playing 1080p+, none of the quad cores (doesn't matter if AMD or Intel) was miles ahead. So I would go for the Phenom II X4 960T BE for about 100€ or the Phenom II X6 1055T 95W for about 140€. It still isn't clear if Piledriver cores will come on AM3+, so I wouldn't waste too much money on the board either.
I was debating the one I've bolded. That will drop straight into my current rig and buy me a little time to decide what to do next. That's probably the option I'll take given that I don't really want to change to Intel but it seems AMD are a little unsteady at the moment.

Thanks for the input!
 
When is AMD going to announce ARM based APU's for mobile?
Never? The ARM landscape is more volatile than the x86/x64 one and they'd need to start from zero wrt to the tight integration they are trying to achieve with their APU line. AMD following up on Intel adding ARM emulation is a possibility though.

I think AMD's current approach in the mobile space is the correct one as far as the APUs are concerned. Still can't fathom how Intel thinks MacBook wannabes at a higher price than the originals are able to crave out a significant part of the market.
 
I was debating the one I've bolded. That will drop straight into my current rig and buy me a little time to decide what to do next. That's probably the option I'll take given that I don't really want to change to Intel but it seems AMD are a little unsteady at the moment.

Thanks for the input!
I'd hold off on a 1055. What specific CPU, and motherboard do you currently have? How soon are you looking to upgrade, and can you be patient (give a rough idea of how long), if it allows for better options? Single GPU? Hi res? Multi-display?

For pure gaming purposes, Intel still wins out. Still, staying AMD may be a better price:performance choice for you, depending on what you have, and what your future intentions are. If the choice is Intel, I have some compiled Ivy Bridge info you may be interested in.


I read an article about a month ago about how the CPU doesn't matter for current games when playing 1080p+
While the GPU tends to be your primary bottleneck at higher resolutions, it depends on how the software you're running is coded, how many GPUs you have, whether physics/other techniques are offloaded to the CPU (think Skyrim), and other factors.


none of the quad cores (doesn't matter if AMD or Intel) was miles ahead.
Generally true, in many instances, though there are a growing number of exceptions.


It still isn't clear if Piledriver cores will come on AM3+, so I wouldn't waste too much money on the board either.
Trinity Piledriver won't be AM3+, as it's a Fusion APU, and has been slated for the new FM2 socket. The FX-series Vishera Piledriver CPUs (what most people here would care about) are going to be AM3+, as that socket was given a stay of execution last year, extending its life. It's expected that both desktop APUs, and desktop CPUs will eventually live on one common socket (FM), with the only differentiations being in which chipsets are available.


Special News: AMD, FX processors are preparing for the 1090FX chip set
http://www.donanimhaber.com/anakart...ciler-icin-1090FX-yonga-setini-hazirliyor.htm

1090FX, and 1070 AM3+ chipsets

- SB1060 south bridge (replacing SB850)
- Up eight 6Gbps SATA ports
- Native USB 3.0 support
- Will supposedly feature further improved VRM design, with higher than current 8+2 phase delivery




It's still a pretty significant jump in performance. You don't get 4-7% CPU performance jumps through a software update every day.

What I'm much more interested in is seeing the processor rematched against Intel's offerings to see where they really stack up now.
I doubt most sites are going to bother with full test suites until the FX-8140, FX-8170, and rest of the B3 steppings arrive.

It is surprising to see so many dismiss an uptick in performance, forgetting that, not only does this update reduce power consumption*, it's also just the first step in a long line of optimizations. For one, it simply treats Bulldozers like SMT, Hyperthreading-enabled Intels. This is fine in many cases, but Windows still needs further optimization to intelligently address that, and scheduling which workloads/threads best benefit from AMD's CMT approach. Not an easy thing to implement.



*I'll post some professional reviews, and user accounts on reduced power consumption.
 
Here's one on consumption, with very early performance figures, as well.

Two things to keep in mind:
1) This is using an un-patched Windows 7, against an early, pre-Beta, Dev version of Windows 8 (article written October 2011).
2) Users with overclocks are reporting even higher power draw savings.


*everything below is a direct quote (including bold portions), with no additional commentary


Sneak Peek: AMD’s Bulldozer Architecture On Windows 8
http://www.tomshardware.com/reviews/fx-8150-zambezi-bulldozer-990fx,3043-23.html
12:00 AM - October 12, 2011 by Chris Angelini

It’s far, far too early to pass judgment on Bulldozer’s performance in Windows 8. However, we already know that Windows 7 isn’t explicitly optimized for the idea of a Bulldozer module. I’ve already talked to the guy from Microsoft who owns process management and threading subsystem in Windows (Arun Kishan), who says Windows schedules to them equally (even if that’s not optimal). And we have a good idea that the next iteration of Windows might be handling them differently.

According to AMD, Windows 8 will more intelligently align threads so that, when they can benefit from sharing a module, they will. The implication is that when two threads can be consolidated onto one module (despite the fact that they’re forced to share resources), putting an entire module to sleep and potentially enabling a higher p-state (a faster Turbo Core setting) outweighs any performance penalty tied to sharing.

I wanted to get a more comprehensive suite of tests run on Windows 8, but the developer build was having none of it. Many of our tests are scripted, and none of those would work. Some of the benchmarks that aren’t (like 3ds Max 2012) wouldn’t install at all.

SolidWorks 2010, Premiere Pro CS 5.5, and World of Warcraft all cooperated, though.



Both SolidWorks and Premiere Pro fully utilize available processing resources. The latter demonstrates no change whatsoever, scoring the same on Windows 7 and Windows 8. The former, on the other hand, ends up favoring Intel’s Core i5-2500K, shaving off eight seconds in the move to Windows 8, as the FX-8150 slows down a little.

The thing is, I wouldn’t expect to see a predictable gain in an application already exploiting available processing resources. All cores are already active; shuffling threads around won't make it any easier to jump to P0. It’s much more plausible to anticipate a speed-up in less-intensive apps, where threads are combined, modules are put to sleep, and Turbo Core is ramped up.



To that point, WoW performance increases measurably on the FX at 1680x1050, while the already-strong Core i5 takes a hit (not something we’d expect from a final build, so perhaps that's overhead-related).

The results at 2560x1600 show Intel’s i5-2500K matching its Windows 7 numbers. Meanwhile, the FX-8150 makes up some of the ground lost to Intel in Windows 7.

Core Parking: Works!

Better still, core parking, the scheduling feature that loads tasks onto as few cores as possible, shutting down the rest to save power, does recognize and optimize for Bulldozer modules. In Windows 7, you’d never see a parked core on the FX-8150. Windows 8 rectifies this, and parks cores in pairs, which we interpret to be Bulldozer modules going to sleep.

Here’s one of example of what you see in Windows 7 (first shot), one example of Windows 8 parking two Bulldozer modules (second), and another example of three modules shut down (third).

The result is a measurable power savings. Idle system consumption under Windows 7 was 107 W. In Windows 8, with three modules parked, that number drops down to 99 W. That's a scant 9 W above Sandy Bridge, a 95 W part. Not too shabby for a 125 W challenger.

Windows 8 Hold Promise

Look, don’t expect Windows 8 to drastically change the Bulldozer architecture’s performance profile. Software is an important part of any hardware review, though, and the upcoming operating system will likely improve the results of some tests once it's finalized. It’ll almost assuredly influence power use more tangibly, as its scheduler learns how to address AMD’s design decisions.
 


Before and After Microsoft's Bulldozer Scheduler Patches: AMD Opteron 'Interlagos' 6274 Server
http://vr-zone.com/articles/before-...amd-opteron-interlagos-6274-server/14572.html

Last week during the CES period, Microsoft finally released a set of Windows 7/2008 R2 operating system patches (KB2646060 and KB2645594) to change core parking and scheduler behaviour to more effectively utilize the "Bulldozer" modules (first dreamt of in 2006) found in newer AMD FX desktop and Opteron server chips. Using a 2P Opteron 'Interlagos' 6274 which is 115W TDP part and has a nominal frequency of 2.2GHz with all cores loaded, we do a quick check for any performance improvements.
Test Setup:

Benchmarks:

.01% to 14% uptick.
 
Thanks feist for the constant updates. Those patch results are pretty decent, can't wait to see how Piledriver shapes up. Perhaps by then the software and hardware sides will be in sync.
 
Thanks feist for the constant updates. Those patch results are pretty decent, can't wait to see how Piledriver shapes up. Perhaps by then the software and hardware sides will be in sync.
You're welcome. Well, along with the added instructions, there are at least two arch improvements that should raise single thread performance. Trinity should offer a nice preview of what to generally expect from Piledriver, later in the year (though PD CPU will have further tweaks over the PD cores in Trinity APU). Sometime after that, Bulldozer 2 arrives, as Intel will already be onto, or beyond, Broadwell ("Rockwell").

Needless to say, AMD have to be on their game.





January 23, 2012 04:01 ET
AMD Opteron™ 6200 Series Processor Family Wins the Linley Group Analysts' Choice Award for Best Server Processor
http://www.marketwire.com/press-rel...alysts-choice-award-best-nyse-amd-1609433.htm

Honor From Leading Analyst Firm One of Several Recent Accolades for the Server Price-Performance Leader

SUNNYVALE, CA--(Marketwire - Jan 23, 2012) - AMD (NYSE: AMD) today announced that its AMD Opteron 6200 Series processor family, based on the new "Bulldozer" architecture, was chosen as the winner of The Linley Group's first annual Analysts' Choice Awards. The awards recognize the top processor products of 2011 in several major categories including embedded, mobile, PC, server, design IP and related technology. To choose each winner, the Linley Group's team of technology analysts focused on merits of the leading products that began shipping between January 1, 2011 and December 31, 2011.

"Our analysts are deeply familiar with processor products and technology, having conducted extensive research in order to deliver the firm's popular publications and reports," said Linley Gwennap, founder and principal analyst at The Linley Group. "We chose the winners on the basis of their performance, power, features and cost as appropriate for their target applications."

This latest award for the AMD Opteron processor comes on the heels of the recent Technology of the Year honors from InfoWorld. AMD Opteron 6200 Series processors along with the Dell PowerEdge R715, which is based on the latest AMD Opteron processors, have both received 2012 Technology of Year awards from the publication.

more...
 
Post-patch benchmarks from Tom's Hardware:

With all of that said, the point of this exercise was to test for improvements in AMD's most recent architecture after several months of patches to Windows 7 and firmware updates. Unfortunately, in our benchmark suite, which is largely made up of applications that our audience has requested over the past year, the speed-ups we were hoping for were not apparent. Only a single application benefited noticeably as a result of the HPC mode that prevents frequency drops in heavy workloads. Thus, most of the discussion surrounding scheduling optimizations remains theoretical, rather than practical.
http://www.tomshardware.com/reviews/windows-7-hotfix-bulldozer-performance,3119.html
 
Post-patch benchmarks from Tom's Hardware:

With all of that said, the point of this exercise was to test for improvements in AMD's most recent architecture after several months of patches to Windows 7 and firmware updates. Unfortunately, in our benchmark suite, which is largely made up of applications that our audience has requested over the past year, the speed-ups we were hoping for were not apparent. Only a single application benefited noticeably as a result of the HPC mode that prevents frequency drops in heavy workloads. Thus, most of the discussion surrounding scheduling optimizations remains theoretical, rather than practical.
http://www.tomshardware.com/reviews/windows-7-hotfix-bulldozer-performance,3119.html
Not sure how they arrived at that, as there is enough evidence to the contrary. Even their own tests - both this one, and the previous Win 8 dev - have shown both performance increases, and reduced power consumption.



That's to go along with other reviews showing increased performance, and greater reduction in consumption as overclocks increased. Plus, better GPU usage (single, and multi-card), efficiency, less thrashing, and smoother performance - even in cases where the actual framerate/time difference is negligible.

Gaining ~10-12% performance increase in Crysis 2, and Left 4 Dead 2 is hardly "theoretical."

Either way, added to the OP, and thanks for posting it.
 
AMD Bulldozer With Improved Performance Patch Windows 7
http://www.jagatreview.com/2012/01/peningkatan-performa-amd-bulldozer-dengan-patch-windows-7/

WEB TRANSLATION said:
Then, whether these improvements can boost the performance of Bulldozer? If yes, how much improvement is generated? Let's look together in the following test.
125w AMD FX-8120
Gigabyte 990FX UD7
Kingston 2x KHX1600C9D381K2/4GX
NVIDIA GTX 560 ti
Coolermaster Silent Pro Gold 800 Watt
Windows 7 64-bit


0% to 7% up


Single-thread application ~ -.04% down


Less than 1% up


~11% to 12% up


~2% up


Wash


Under 1% down, to 3% up
WEB TRANSLATION said:
Analysis

Comparing conditions before and after using the patches made we found some interesting results. Overall performance improvements seen after the use of the patch. Some applications on the list of testing showed increased high enough, the other did exist even though the increase is not too large.

Common software performance indicates that this patch produces the greatest increase in applications that use two threads. On the use of other threads visible improvement is not too big or nothing at all.

Gaming performance showed significant results when the games are played optimized for dual core performance. At the time the game is using all the threads that have processor performance is also increased although not as big as when using a dual thread.
WEB TRANSLATION said:
CONCLUSION

Having seen the use of patch testing to improve the performance of the Bulldozer. The increase was most visible at the time of application that uses the thread a little, when using a visible increase in the maximum thread that is not too large.
 
Can anyone give a guess as to how the desktop variant of Trinity will compare with my Phenom II/4850? Im looking for maybe a 10% upgrade to play mainly Diablo 3 and SC2 expansions.

Ill probably go with a i5/7770 but the thought of a capable mini-atx machine is equally enticing.