McHuj
Member
What's going on here? Same specs, same stepping, different TDP?
Probably refinement in the manufacturing process. As stated above 32nm is getting better and better.
What's going on here? Same specs, same stepping, different TDP?
Probably refinement in the manufacturing process. As stated above 32nm is getting better and better.
It gives AMD flexibility. The 95W 8100, and 95W version of the 8120 are OEM specials, with the 125W 8120 being in wide availability. Outside of that, retail 95W 8120s are unicorns. Remains to be seen how they handle the 8140.
What's going on here? Same specs, same stepping, different TDP?
Not necessarily. It's largely in how they're binned.Wouldn't that mean a different stepping, though?
WEB TRANSLATION said:While Intel is occupying the higher performance classes AMD have a golden opportunity to attack the broad mid-range. Thanks to the well-oiled AMD graphics division can transfer the technology from discrete graphics processors to the same silicon as the traditional processor. Next up is the second generation A-series, known as the Trinity.
According to Sweclockers released the first processors based on Trinity May 15. They are starting with models of desktop computers and traditional laptop. Extra power-optimised variants for ultra-portable computers will take until June.
AMD Trinity features two to four processor cores with Bulldozer architecture, known as the Piledriver. Computational units are flanked by a new video part from generation DX11. According to AMD, the latter up to 56 percent higher performance than today's A-series.
AMD Client Overview - CeBit 2012
AMD's coming Trinity architecture is not far off, and has now been benchmarked in Folding with a mobile version: A8-4500M. The processor performance is actually better than expected, even though we are dealing with Piledriver cores.
Trinity is AMD's next generation APU, and comes with two Piledriver module, which results in a total of four processor cores. On the same silicon we find six VLIW4 clusters that build on the same architecture we find in the HD 6900 series, but many of the functions found in the new HD 7000 series has also made it. It is therefore expected by AMD to deliver on the integrated graphics and raise the bar even higher on that part.
Piledriver is an evolution of Bulldozer, which has made many question its processor performance, especially after the first benchmarks leaked. What we believe is a manufacturer has tested the mobile version of Trinity in Folding and the benchmark show promise. Something that speaks in AMD's favor is the high clock frequencies that can compensate for the lower performance-per-clock with the new architecture. The benchmarks are with two different systems and the results in Folding varies.
The two cores in a Piledriver module shares one floating point unit, so the boost there is relatively marginal. When it comes to integer performance the boost is nearly 40 percent, this is even though the new processors are specified to the same TDP. AMD told us we can expect up to 25 percent better processor performance with the same energy consumption over last generation, which matches the results above. The exact frequency of A8-4500M is uncertain at this time, but should be higher than 1,5 GHz found with A8-3500M.
Later this year AMD is going to introduce "Trinity" processors, that utilize enhanced Bulldozer micro-architecture, codenamed "Piledriver". New processors will be released for both desktop and mobile markets, and we already know specifications of several desktop Trinity parts. First details of mobile Trinity CPUs were revealed in February, when Arctic announced MC101 HTPC. The Arctic MC101 system will come with quad-core A8-4500M and A10-4600M SKUs, which integrate HD 7640G and HD 7660G graphics. The frequencies of these quad-core CPUs were not disclosed. Third Trinity-based mobile microprocessor recently appeared in specifications of HP g4-2002ax and g4-2003ax notebooks. The new part has model number A6-4400M.
Based on high clock frequency, we suspect that the A6-4400M CPU will fit into 35 Watt thermal envelope. Other CPU features, such as the size of on-chip caches, are unknown.
With the [Windows 7 Bulldozer] updates in hand, today we are looking at five 990FX boards that may feature on the consumer or system builders radar. This roundup has been on the cards for a long time, but unfortunately has had to be continually pushed back and then retests applied with latest BIOS updates but as belated as it might be (and as deeply apologetic as I am), here it is!
AMD has notified its partners that some new parts are already in production.
The most notable new processor is the A10 5800K that went out of the production sample phase in early February and it went into production in early March. A10 5700 was in production sample phase in mid-February and it entered production in mid March. A8 5600K is in production since early March as well, while the A8 5500 is in production as of mid of last month.
A6 5400K and A4 5300 should enter production ready sample phase in early April and move to full scale production in April 2012. The current schedule is that A10 5800K, A10 5700, A8 5600K, A8 5500 start shipping in Q2 2012 while the last two A6 5400K and A4 5300 start shipping in Q3 2012.
The prototype laptop AMD had brought along with them had undisclosed specs, but they confirmed that it was running their top of the line A10 quad-core APU. This test unit specifically was running just the APU, there was no discrete graphics card within the laptop, ergo no CrossFire.
While they didnt allow us to run the built-in benchmark on Dirt 3, we did get a chance to see the game running on the A10 APU with everything set on high and 2xMSAA. Check it out for yourself below.
Next up is a short demonstration of one of the new features AMD is implementing in the A10 APU. The AMD Steady Video technology is implemented at a hardware level on all Trinity based APUs, and one of the videos AMD demonstrated this image stabilization tech with was the extremely shaky Triple Tomcat Launch.
Lastly we wrap up the A10 Trinity demo with a short interview with AMDs Stefano Chiavegati, Head of Commercial and Component Marketing for EMEA. We asked about the A10 Trinitys launch, comparison with Ivy Bridge, CrossFire compatibility with HD 7600 series cards.
So expect the AMD A10 Trinity APUs to be out by the end of June. Well have more updated coverage on the new platform closer to release.
Since February I'm regularly searching for appearances of new "family 21 model 16" BOINC results, which belong to AMD's Trinity APU. As I noticed, I'm not the only one doing that. ;-) Some early results of an engineering sample (ZD372058A4451_41/37/16_9901_800, which should clock at 3.7 GHz base and 4.1 GHz turbo clock according to the string) didn't look bad (one day it reached an integer score of over 13K on 64b linux). But to do some halfway accurate (or semiaccurate ;-)) analysis it is important to look at results achieved on the same OS (here: Win 7, 64 bit) and BOINC client version (6.12.34 here except for the ES, which run a 6.12.43 client).
The FP benchmark, which is a Whetstone benchmark, seems to run as a multithreaded benchmark according to "informal". At least it fills up all available cores while running. The integer benchmark, a good old Dhrystone benchmark, seems to be single threaded. Further it is important to know, that both benchmarks have a rather small memory footprint.
Since we don't know the exact clock frequencies of the benchmark runs, it is difficult to find the correct value for calculating per GHz results. I estimated those based on turbo clocks, which might lead to skewed results. At least in the case of comparing Trinity with its Piledriver cores to the FX models, I hope that rather similar turbo mode behaviour should reduce the error margin.
OK, here comes the table comparing several values I filtered out of my collected BOINC results to have OS and client version the same. As you can see, Piledriver w/o L3 cache seems to perform a bit better than BDver1 based FX models:
Note: I used "Trinity vs. Bulldozer" to denote the difference between a L3-less Piledriver core and a Bulldozer core, which always had L3 available.
A few months ago during the Radeon HD 7970 launch we discussed how AMD would be revising their ISV relations. While their efforts with ISVs in the past few years had been decent, AMD had not put a lot of money into it, and what money they did have was controlled by a relatively large bureaucracy. Consequently their efforts paled in comparison to NVIDIAs, who put in far more money and effort into the process.
As a result one of AMDs reforms for 2012 was going to be that they were going to put more money into ISV relations in order to catch up to NVIDIA. And while our discussion with the 7970 last year focused on the gaming side of things, AMD is also throwing more money behind ISV relations for their professional graphics business. After hitting a significant low against NVIDIAs Quadro lineup, AMDs market share for professional graphics has rebounded based on the strength of their 40nm DX11 GPUs, and AMD is looking to keep that momentum going with these improved ISV relations.
With Creo Parametric 2.0, AMD and PTC have added support for a couple new features intended to significantly boost performance.
With somewhat disappointing Bulldozer showing, and perceived wavering roadmap, there were comments in the industry circles how AMD will give up the main Intel competitor seat to someone like Qualcomm, or even cease to be in the CPU business altogether. This weeks' preview of the next-generation Fusion APU gave us a chance to mingle with AMD's top execs, see their new CEO and hear - unofficially, of course - what's the way forward.
The hot, dry Austin, in Texas, is the place where the top AMD management is, after all, so the chance of meeting high-level personnel at a press & analyst event is high. This time we're not talking more about Trinity yet, except that the thing does seem to perform well, and yes, there is a 17 W Ultrathin-class (read: UltraBook in Intel speak) Trinity with all the features enabled, i.e. two dual core modules and the full GPU - of course at a lower clock speed. That will make some interesting impact in that market later this spring...
The APUs aren't a problem either: the Trinity is expected to be sold out, whatever AMD can make - so, we come to the quantities question there. AMD gave assurances they can do many millions of Trinities this year, and, according to what was seen, the 'good enough' statement, as worn as it sounds, is right on target here.
Then we come to the CPUs, the real problem for years - I did overhear a discussion between the senior execs how did the whole thing turn out with the 'low IPC performance' Bulldozer instead of the other alternatives on the table, and how it will take some two years to completely turn around the switch to another much more IPC-efficient core architecture. The main culprit seems not to be the previous CEO, Dirk Meyer, but his predecessor Hector Ruiz, the 'Taco Bell' that replaced 'KFC goatee' Richard Sanders, the man who made AMD succesful.
They said that Sanders was so disappointed with Hector, saying once that 'he left Hector a roadmap to execute that anyone, absolutely anyone' would be able to get done. Of course, Hector did damage the company in many ways, which by itself is a separate story.
Keep in mind, by then Intel will have worked out most, if not all, of the kinks with Tri-Gate. Broadwell will already be online (big GPGPU iGP advancements and all) and Sky____ in the pipe.Talking about moving forward, the feel is that AMD is there to survive and thrive once again. Rather than repeating the bland statements of 'we don't want to compete in the high end, just on volume' which make no sense since a fabless company cannot compete on volumes against a competitor with seven large fabs at its disposal, this time there's clear indication that, after Piledriver, there will be substantial changes in both cores and system architecture from Steamroller onwards, that should help make AMD competitive closer to the top. I was told that delaying the socket migration beyond the AM3+, C32 and G34 to new socket is a good move, since AMD can design more aggressive, rather than stop gap, sockets for future platforms with better features like more memory and HyperTransport channels, as well as integrated PCIe v3, for greater future scalability. For the first time, some execs do acknowledge that Bulldozer approach may not have been the best one at the time, and things need to change. I was told that there is some good frequency scalability in the Piledriver core which should help gain some per-core performance ground.
So, in the near term, AMD will use APU to keep its presence in desktop and mobile market, and even low power 5 - 10 W part derivatives or Trinity may arrive for high-end HD++ tablets. The CPU core radical refresh is expected to complete within two years from now, along with brand new socket platforms, proving AMD a new base from which to attack the high end, again, just like in the good old Opteron/Althon 64 early times. The GPUs will continue to be the crown jewel of the company till then, though... expect new high end mobile HD7900 series later this month, and Sea Islands by yearend.
GPU - hd7660g vs hd6550d vs hd3000 vs hd6620g
2D
3D
WEB TRANSLATION said:Integer math and find prime, numbers seems to be a bug led to scores improved a lot
One of 3500m integer and find prime numbers than those other exceptions of many
This week, renowned tech media around the world (including HardwareZone) were given intimate insights showcasing the capabilities of AMD's Trinity APU and the nuts and bolts of its design.
While all the technical details, features and performance results are under embargo for now and youll have to wait till 15th May for updates, what AMD has allowed to share at the moment is the experience of the AMD Trinity based notebook. Understanding that this class of products with the AMD Trinity APU are targeted at mainstream users, the company focused on improving end-user experience rather than focusing on raw benchmark numbers. A combination of the reasonably powerful graphics engine on the APU and accelerating applications enhanced with OpenCL API which AMD has been strongly pushing the software community in the background, they are able to improve what matters most for notebook users - the everyday tasks.
To showcase this, AMD setup a blind challenge with systems hidden away and the tech journalists are only faced with identical monitors and a custom-built start button to get the comparisons started at the same time on both an AMD Trinity based notebook and a somewhat identically configured Intel notebook. With three different scenarios touching upon office productivity, home video playback and file compression, here are the results in brief from gathering feedback from all the journalists who participated:-
The results say it all that most participants (including ourselves) preferred the AMD Trinity system's experience. Now to put some perspective, these are findings from just a small sample of benchmarks used, of which most were running updated software benefiting from AMD's technologies or taking advantage of its built-in discrete-class graphics engine via the OpenCL API. While it may give AMD the upper hand, it also gives us an idea of the possibilities with the new AMD platform when the full set of features are exploited and put into good use.
So far it's looking positive as we found office productivity tasks loading and interacting a little faster and the same goes for file compression handling. For the video playback, it used a typical amateur home video clip and we noticed the AMD system applying post processing effects to improve color accuracy, remove noise and with AMD Steady Video technology, it even removed all the jerks and shakes normally associated with home videos. The Intel system on the other hand just played the video as-is. While the upcoming Intel Ivy Bridge processors could pose a tougher challenge, most are confident in AMD's lead where graphics are concerned because of their experience in better hardware and software.
While the Trinity APU is officially launching on the 15th May, AMD expects more notebook offerings to be available from June onwards. The desktop counterparts will come in later in the third quarter of the year.
The green bars look so exaggerated in this chart, lol.
A8-3500m: 1.5GHz, 2.4GHz turbo 4C/4T
i5 2410m: 2.3GHz, 2.9GHz turbo 2C/4T
A10-4600m: 2.3GHz, 3.2GHz turbo 2CU/4C/4T
According to our sources, it looks like AMD is not done with acquisitions, and their second potential acquisition might be the second most important one they did in 20 years. While it is too early to tell were our sources misleading us (thus, take this piece with a fair grain of sea salt), the story was told to us with compeling arguments.
Echos of NexGen
[...]
MIPS: An Amazing Turnover
[...]
During a panel with senior executives from Linley Group, Intel and Qualcomm, Amit Rohatagi, Principal Architect for MIPS Technologies detailed the mobile strategy of its company and the effects of collaboration with Chinese Ingenic Semiconductor. Ingenic was founded in 2007 and until today, the company shipped over 30 million CPUs. Ingenic is focusing on optimizing MIPS architecture and is currently working on new mobile chip based on MIPS64, a 64-bit architecture which debuted in 1991.
The growth of MIPS in tablet space is especially interesting. Since the announcement that Google Android 4.0 ICS is enabled to work on MIPS processors, MIPS started to grow in Mainland China like no tomorrow. Chinese Ainovo (products are sold under Ainol brand) was the first manufacturer to offer an 7" Android 4.0 ICS powered tablet for only $99, and after the company offered $79 and $119 parts, all tablets are continuously sold out. Ainol is producing as much tablets as they can handle, and the company is now starting to see competing companies launching their MIPS powered tablets.
In addition, ARM, among many others, are currently licensing IP from MIPS.The numbers are quite impressive - for a part that only hit mass production at the tail end of last year, Ingenic Semiconductor shipped 1.8 million chips for tablets. With the initial programme being deemed successful, the next step is to push for 10 million shipped units by the end of the year.
Given that even Philips launched a low-cost tablet product based on MIPS architecture (available in China only), time will tell when the world are going to take notice. Remember, just six months ago, nobody expected MIPS to emerge as a serious contender to ARM and x86.
Forbes, for one, its ears to the ground, is betting on the AMD rumor, pointing out that "MIPS has a chip that runs Android 4.0 ICS and consumes LESS power than competing ARM parts. Furthermore, that chip is built in old 65nm process versus power-saving 40nm and 28nm competitors."
Forbes thinks it would be a lot smarter for AMD to buy its own chipmaker for its new microserver acquisition than to license ARM as widely expected.
It might also replace some of the engineering talent AMD has lost.
MIPS is supposed to sample a two-core dingus that targets the ARM A9 later this year but again consumes less power and it's got 64-bit chip targeting the A15 in design. ARM is supposedly having trouble with 64-bits.
[...]
http://www.anandtech.com/show/5865/laptop-graphics-face-off-diablo-iii-performance/3
1 - Hot Chips 2012: An Introduction to Surround Computing
2 - Steamroller in Slightly Better Focus
1 - Front End & Execution Improvements
2 - Cache Improvements & Looking Forward
So, yeah, no 2x256-bit FMAC (shouldn't have been expected), but those who doubted the decode increase were clearly mistaken.
At Hot Chips today, AMD is revealing some of the details about their new Jaguar core, the successor to Bobcat, the first of their smaller ‘Cat’ family of cores. Take four of these second generation low power cores, add a GPU, and off you go to power lower end laptops and higher end tablets.
Those among you that follow chips closely might realize that Jaguar is not the second iteration of the smaller ‘Cat’ family of CPUs and APUs, that was supposed to be the Witchita and Krishna twins. As SemiAccurate exclusively pointed out last November, those fraternal twins were canned for a number of reasons and the Jaguar cored Kabini was pulled in. Today AMD starts the long process of revealing this chips, but things like the speeds, die areas, and most of the uncore were not revealed. In fact, AMD didn’t even mention that it is a 28nm part.
So what did they talk about?
Sections:
- The Jaguar core; color coded by function and not to scale
- Pipes and more pipes, by the clock
- Bobcat vs Jaguar IPC and power gating
- The cores, new and old, portable and more portable
AMD FX-Series Vishera Specifications Found
The FX-4320 is the entry-level CPU, insofar as a quad-core with a clock speed of 4.0 GHz (4.2 GHz Turbo Core) can be considered low-end. Its TDP (thermal design power) is of 95W and the cache memory (L3) is of 4 MB.
The middle CPU is called FX-6300 and has six cores. The base clock speed is 3.5 GHz while the Turbo Core frequency can go to 4.1 GHz. The TDP is 95W and the cache is 6 MB.
And, finally, we reach the flagship Vishera FX Piledriver Bulldozer, the 8-core FX-8350, with a stock speed of 4 GHz and a Turbo Core setting of 4.2 GHz.
standard air:
extreme cooling:
Table of contents
01. The Story Of Fusion Begins
02. Looking For The Other Half
03. Merger And Mayhem
04. Scaling The Brick Wall
05. Up From The Ashes
06. Fusion Ignites
07. Heterogeneous Roots
08. OpenCL And HSA
09. Focus On The Programmer
10. HSA's Big Picture
11. More About The Big Picture
12. HSA Tomorrow
It really is.Feist needs a cohort he can talk to in this thread. Or even a nemesis he can argue with.
It's an impressive collection of info you've created in this thread nonetheless.
It really is.
And it's hard to get excited about any of it.
Hope AMD can do anything to compete with Intel in the enthusiast desktop market.
*And the CPU/GPU it looks to be ahead on which was a good focus a while ago. I'd easily recommend a system to some normal person that has that. Onboard video and even the HD2000 feels crummy.
Thanks. With all the CPU/GPU talent they've lost to Samsung and Qualcomm, getting Gustafson and bringing Jim Keller back into the fold are both good moves.Nice having you back here. =)
Intel's John Gustafson, known for Gustafson's Law, just joined AMD. It's clear that AMD is aware that it's the software that matters. I just wish existing software could always magically make use of current silicon improvements.
Same, it's a little underwhelming.
Maybe they're ahead of the curve and expected more games and apps to be better threaded, but they aren't.
And seeing terms like 'refine' or 'improve' alongside things like 'surround' computing feels like they're going for the lower end embedded systems rather than power. Even in laptops Ivy bridge brings HD4000 which is a nice step up in power (and both intel and AMD integrated solutions are still lacking for all but the most casual gamer)
Don't want it in my next console particularly either..
What's weird is it seems like developers want to take advantage of the gpu muscle more than they want to optimize code for more cores.
I wish more articles would be worth wading through all the text. AMD's current relative strength is the total system efficiency (power per watt of CPU, GPU and NB) while keeping in mind, that it still runs on the bigger 32nm silicon. Most articles gloss over it and at best take AMD's TDP which is split in set groups (9, 17, 18, 25, 35, 45, 65, 80, 95, 100, 125W...) and reflect the worst case wastage allowed, not the actual average energy efficiency. I guess that particular approach fired back big time with the press simply comparing that with Intel's different use of TDP.
How is the developers' focus on graphics weird? That's more like a systemic plague. Changing capabilities on the GPU side are accepted both more and for a longer time so the reluctance to make the comparably small step to make use of it also in GPGPU ways is naturally lower.
Some further AMD news:
With CodeXL AMD now offers development tools for both Windows and Linux for profiling CPU and GPU, debug GPU and analyze static OpenCL kernels.
For Windows AMD has an own app store called AMD AppZone which now has been extended with Android apps which run using the Blue Stacks Player, a (supposedly hardware accelerated) ARM emulator AMD invested in.
What's weird is it seems like developers want to take advantage of the gpu muscle more than they want to optimize code for more cores.
Not at all, really. Fine grained optimizations are cost-, and time-consuming. Most companies will do a cost-benefit, and opt to push products to market in different states of refinement. It's far easier to rely on brute force compute resources, as they always have. Now, with CPU and process mnfr advances increasingly coming at a premium, it's only natural to opt for GPGPU, and the various "easy" performance initiatives such as HSA, or AVX2.What's weird is it seems like developers want to take advantage of the gpu muscle more than they want to optimize code for more cores.
Yeah. GPUs are vastly more capable in most non-serialized tasks, and even the bad ones are still float point beasts. In FP64 Kaveri's GCN iGPU is a considerable step forward from Trinity's VLIW4. It's a nice GPGPU part, bandwidth allowing. I'm oversimplifying, but oddly enough, Intel and AMD's current approaches to iGPU performance are near inverses of their CPUs. Intel is largely throwing more execution resources at the top end Haswell, while AMD is going with higher "per-core" efficiency with Kaveri, in addition to more resources. Things begin to get truly interesting when the successors to Kaveri and Haswell arrive for 2014 and ~2015, respectively.Not sure i got this right but arent GPU not more bang/watt processors. Only got a brief introduction of the subject at current internship.
Wait, you were expecting tech sites to do proper performance-per-watt testing, along with real world usage in varying loads and tasks? Oh, you. People can hardly overclock correctly. Even the most respected sites can have glaring omissions, odd methodology, and lapses in basic level understanding of tech.I wish more articles would be worth wading through all the text. AMD's current relative strength is the total system efficiency (power per watt of CPU, GPU and NB) while keeping in mind, that it still runs on the bigger 32nm silicon. Most articles gloss over it and at best take AMD's TDP which is split in set groups (9, 17, 18, 25, 35, 45, 65, 80, 95, 100, 125W...) and reflect the worst case wastage allowed, not the actual average energy efficiency. I guess that particular approach fired back big time with the press simply comparing that with Intel's different use of TDP.
·feist·;42865950 said:Wait, you were expecting tech sites to do proper performance-per-watt testing, along with real world usage in varying loads and tasks? Oh, you.
NordicHardware has seen exclusive information about a new energy law that will apply within the EU. The law requires that both discrete and integrated graphics cards live up to certain energy standards. AMD is worried that this will affect next generation graphics cards and have them barred from sales in the EU.
The Vishera review emargo is apparently up since I'm seeing a few stories going around. MaxPC just posted a comparison feature about how it stacks up to Ivy Bridge performance wise.