• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

3D Mark releases new DX 12 benchmark - TimeSpy

dr_rus

dr_rus

Member
Baron_Calamity said:
How so? My CPU score is half a good number of other scores posted. Even to some cpus below its class. The i7 3770 is a good cpu but mine is slower than most 3770 for some reason and even if it wasn't, its showing its age.
Click to expand...

You're unlikely to see many gains in actual games with a CPU upgrade while with a new videocard you can easily get 2x+ framerates even with the same CPU. 3DMark's CPU test is pure synthetics, no game in the world is using the CPU like that test.
 
Pjsprojects

Pjsprojects

Member
After trying over twenty overclock setups I noticed I was loosing up to 200 in the total score output as time passed.
The reason is cpu heat,I'm using a 212evo cooler and hitting 55 degrees,where as a mate has one of the Corsair closed loop coolers and runs at 30 degrees.
I guess those coolers are worth the expense!
 
Larogue

Larogue

Member
GeForce GTX 680
Intel Core i5-3570K Processor
No OC on both
Score 2210

(Graphics Score 2108. CPU Score 3047)

Not bad, considering the fact I'm using a video card form 2012.
 
E

Evo X

Member
Just over 6K with this overclocked Titan X. 6015 to be exact. 5854 GPU and 7138 CPU (5820K)

Can't wait to revisit the benchmark with Pascal Titan to see the improvements in architecture.
 
Hatty

Hatty

Member
UV3Yfzj.png
Click to expand...
My score seems lower than it should. Maybe somethings wrong
 
C

chromatic9

Member
Is there a 1-3gb download just for time spy? Steam wants 7gb for the package, techpowerup has a 4gb version, not sure if it will update to 7gb to add time spy.

3dmark has become too big for me to want to have that on my SSD now and I have a 1tb SSD, don't even want to waste bandwidth of 7gb just to run time spy.
 
Z

ZOONAMI

Junior Member
Highest overall score for xeon1231v3 and 980 ti G1.

http://www.3dmark.com/spy/84145

5696 score and 6,119 graphics score for the G1 980 ti at 1500/8000. Funny that I have 2 rigs that score nearly exactly the same. Other rig is a 1070 and i5 4690k. Pretty amazing for the 1070 given the power draw. 1070 does like 6,300 graphics score.
 
Z

ZOONAMI

Junior Member
Lonely1 said:
Well, looks like this is the best I can do without making my system unstable:

http://www.3dmark.com/3dm/13390915?

Can't break the 6K barrier. For some reason I can't OC my video memory. This means that I lost the silicon lottery?
Click to expand...

Kind of if you can't oc your memory at all. Guessing that you have elpida or hynix memory. Samsung memory can usually do 8000mhz as it's speced for that. My 980 ti G1 does 8000mhz for the memory no problem. Your core clock boost is pretty good though, so I wouldn't say you lost, just that you didn't win.
 
L

Lonely1

Unconfirmed Member
ZOONAMI said:
Kind of if you can't oc your memory at all. Guessing that you have elpida or hynix memory. Samsung memory can usually do 8000mhz as it's speced for that. My 980 ti G1 does 8000mhz for the memory no problem. Your core clock boost is pretty good though, so I wouldn't say you lost, just that you didn't win.
Click to expand...

Yep, GPUz says Hynix. :(
 
dr_rus

dr_rus

Member
A Closer Look at Asynchronous Compute in 3DMark Time Spy

Summary

Asynchronous compute is one of the most interesting new features in DirectX 12. The work that Time Spy places into the COMPUTE queue and the specific implementation of that work is the result of deep co-operation with all BDP members including AMD, Intel, Microsoft, and NVIDIA among others.

The implementation is the same regardless of the underlying hardware. In the benchmark at large, there are no vendor specific optimizations in order to ensure that all hardware performs the same amount of work. This makes benchmark results from all vendors comparable across multiple generations of hardware.

Whether work placed in the COMPUTE queue is executed in parallel or in serial is ultimately the decision of the underlying driver. In DirectX 12, by placing items into a different queue the application is simply stating that it allows execution to take place in parallel - it is not a requirement, nor is there a method for making such a demand. This is similar to traditional multi-threaded programming for the CPU - by creating threads we allow and are prepared for execution to happen simultaneously. It is up to the OS to decide how it distributes the work.
Click to expand...

A lot more at the link, a good read really for those interested to figure out what async compute actually is.

icecold1983 said:
http://www.futuremark.com/pressreleases/a-closer-look-at-asynchronous-compute-in-3dmark-time-spy

press release going into more details to end the flame wars. everything seems pretty fair
Click to expand...

Oops, I was late. Doubtful that this will end the flame wars though 8)
 
I

IC5

Member
Baron_Calamity said:
igfL3b7.png


What can I say?

giphy.gif


I really need a new computer.
Click to expand...

eh, it depends. Its a pretty benchmark, yeah. and its a bummer when your gear can't run it smoothly. But it doesn't really mean anything, in regards to the games you are actually playing.

Your processor is still pretty awesome. Ivy Bridge is a great core. Your 3770k has a long life ahead of it. Even at stock. If you aren't already, overclock it it a bit. 4.4 or 4.5 Ghz should be pretty easy to hit while still getting good temps on a decent air cooler. I suspect overclocking is the main reason you aren't posting as good of numbers, as others. Some of those people are also probably running faster ram, than you are.

It doesn't say exactly what you have. But the 7950 and 7970 are both still solid GPUs. You won't be maxing the latest games anymore, at 1080p. Like you were a couple of years ago. But you will still be doing better framerates than a PS4, with some extra graphics options, on top of that. I have an overclocked 7870 and the only thing I wish for is 2 extra gigs of VRAM. >_>

Here is my score
GPU = 7870 Ghz edition with 1140 core overclock and 1325 VRAM
CPU = i3 6100 Skylake with Hyperthreading enabled
RAM = 16gb of DDR4 3000 15-15-15-35
Adata SP550 480GB SSD.

all crammed into a rebranded Raijintek Metis mini-itx case.
SEVb.jpg
 
LostDonkey

LostDonkey

Member
Baron_Calamity said:
How so? My CPU score is half a good number of other scores posted. Even to some cpus below its class. The i7 3770 is a good cpu but mine is slower than most 3770 for some reason and even if it wasn't, its showing its age.
Click to expand...

Have you checked nothing is running in task manager before you do the bench? My 6700k was getting scores way below others and I had a couple of processes taking up cpu time so I killed them, re ran the bench and got a score on par with others.
 
R

Raticus79

Seek victory, not fairness
Finished my build - still working out some bugs

Edit: good enough for now. This is at 4.2GHz on the CPU and 3200 on the RAM (base XMP profile). Should be able to get a bit more out of it.

44Vz4mQ.png
 
D

Durante

Member
dr_rus said:
A Closer Look at Asynchronous Compute in 3DMark Time Spy
Click to expand...
Whether work placed in the COMPUTE queue is executed in parallel or in serial is ultimately the decision of the underlying driver. In DirectX 12, by placing items into a different queue the application is simply stating that it allows execution to take place in parallel - it is not a requirement, nor is there a method for making such a demand. This is similar to traditional multi-threaded programming for the CPU - by creating threads we allow and are prepared for execution to happen simultaneously. It is up to the OS to decide how it distributes the work.
Click to expand...
You'd hope this much would be obvious for anyone fervently participating in a technical discussion about a renderer (or worse, making accusations).
 
Z

Zexen

Member
2016-07-23_14581948aj5.png


I really need to set up my WC, the proc keeps crashing with my usual OC this summer. Here it's stock clock.
 
L

low-G

Member
Does anyone know if the hyper threading gains shown in this benchmark are at all realistic for a real world DX12 game scenario?

Just the addition of hyperthreading seems to add almost 25% to the CPU test performance. I guess CPU-only game calculations aren't realistic in the first place... Is this just a case of 'poor' thread management being handled better by HT? Because I'm not talking additional cores.
 
L

Lonely1

Unconfirmed Member
low-G said:
Does anyone know if the hyper threading gains shown in this benchmark are at all realistic for a real world DX12 game scenario?

Just the addition of hyperthreading seems to add almost 25% to the CPU test performance. I guess CPU-only game calculations aren't realistic in the first place... Is this just a case of 'poor' thread management being handled better by HT? Because I'm not talking additional cores.
Click to expand...

No current game pushes the CPU as hard this test for obvious reasons. But hyperthreading does improve multi-threading CPU performance, whenever that affects gaming performance will change from game to game basis.

In particular, if games are developed for consoles first and doesn't include a slider for things as "crowd density" there might be little to no difference. I don't know if games are starting to use more the 4 threads either.
 
Celcius

Celcius

°Temp. member
Here are the results from my GTX 780 Ti (stock):

Overall Score: 3025
Graphics Score: 2892
Graphics Test 1: 19.07 fps
Graphics Test 2: 16.42 fps
CPU Test: 4093
CPU Test: 13.75 fps

My card was paired up with a 2600K @ 4.6ghz and 16gb of DDR3 @ 2133mhz.
 
P

Pachinko

Member
I finally ran this today , got 5519 with a 6700k paired with a 1070. No overclocks or anything set up. Pretty neat tech demo.
 
A

AP90

Member
Ran time spy on my 2600k 980ti rig and with having my 980ti sc at default clock settings. It was artifacting in an out and going black throughout the test, then popping back in and still finished with a score of 5459. I recently did a computer wipe and installed a bigger ssd.

Still trying to figure out what the problem is. My system runs 3dmark fire strike, games and the other benchmarks with no problems.

Edit: At some point after work today.

Planning to reinstall the latest NVidia drivers to see if it fixes it and then trying it out on a different monitor..It's almost like the engine is not outputting the visuals to my monitor properly.
 
Arc07

Arc07

Member
My score went down over 1000 points with the latest Nvidia drivers. Anyone else having a similar issue?
 
You must log in or register to reply here.

Similar threads

"I Need a New PC!" 2024. 240 Hz OLEDs, PCI-Gen5, Path Tracing & Ray Reconstruction.
4 5 6 7 8
399
17K
Rossco EZ replied
  • Locked
"I Need a New PC!" 2022. The GPU drought continues...
23 24 25 26 27
1K
74K
DGrayson replied
AMD announces frame generation for all DX11 and DX12 games
2 3 4
169
12K
Crayon replied
"I Need a New PC!" 2021. Sold Out Edition.
24 25 26 27 28
1K
102K
Rex_DX replied
Intel Core i9-10900KF 10 Core, 10th Generation Desktop CPU Benchmarks Leak – ALMOST On Par With AMD’s 12 Core Ryzen 9 3900X But 14nm A Troublesome Bet
News Rumor Hardware
33
3K
RoboFu replied
Top Bottom