Going from 1090T to i7-980 with a 3930k for comparing

[bmadd89]

I have posted a few times with regards to upgrading my editing machine and i just ordered a 3930k and Asus rampage IV Formula with 32GB ram to replace me current 980 @ 4Ghz on an Asus P6T SE with 24GB of ram.

I'm going to be throwing the 980 into my gaming machine which currently has a 1090t @ 4.18Ghz in an Asus crosshair with 8GB of ram.

I know we all know the 980 will lay my 1090t over the kitchen counter for gaming (mostly what this machine does). I'm goal is to illustrate the difference at the same clock speed for all 3 processors. I'm doing this because i can read benchmarks and look at charts all day from review sites but i can never seem to replicate the same results as them, and they don't generally compare with these kinds of OC's. That and i personally want to know what the difference and i bet there would be a few others as well.

I can run a direct comparison between the 1090t and the 980 with the same gpu but i will be limited to just the CPU when i get the 3930k because i have water in the gaming machine.

So i will start to run benchmarks on my AMD system now and when i'm done i'll change my system across to the 980 and retest and then i will add the 3930k cpu scores where appropriate when they arrive. I can also test the CPU benchmarks on my stock 2600k at work if people would like that included as well.

I'll test IBT for 20 loops on standard stress.
I'll test CineBench 11.5 64bit.
I'll test CineBench 10 64bit.
I'll test WPrime 32M
I'll test WPrime 1024m
I'll test 3Dmark11 on performance.
I'll test 7Zip 32m benchmark.
I'll test 3Dmark Vantage on performance
I'll test Crysis Warhead bench on very high
I'll test Stalker: Call Of Pripyat benchmark
I'll test Unigine 3

Is there anything else people would like me to test that is easy and fairly quick to set up?
I'll test POVRAY 3.7 RC6 as requested by jhu

Graphics card. Running 12.7 beta
AMD HD7970: Core @ 1250Mhz. Memory @ 7200Mhz

AMD 1090T @ 4180Mhz. NB @ 2935Mhz. Ram @ 1505Mhz 9-9-9-24. HTT @ 2032Mhz
Intel i7-980 @ 4186Mhz. NB @ 3220Mhz. Ram @ 1610Mhz 9-9-9-24. QPI @ 5796Mhz
Intel i7-3930k @ 4200Mhz. NB @ ???. Ram @ 1333Mhz 9-9-9-24. QPI @ ???

I will compile a spreadsheet of all the data for each benchie once i have completed

COP settings

Crysis Warhead Settings

 

[bmadd89]

AMD 1090T. Speed in first post

IBT
----------------------------
IntelBurnTest v2.53
Created by AgentGOD
----------------------------
Processor: AMD Phenom(tm) II X6 1090T Processor
Clock Speed: 4.18 GHz
Active Physical Cores: 6
Total System Memory: 8190 MB
Stress Level: Standard (1024 MB)
Testing started on 10/07/2012 10:38:47 PM
Time (s) Speed (GFlops) Result
[22:39:14] 11.568 77.2703 3.257006e-002
[22:39:40] 11.504 77.7000 3.257006e-002
[22:40:06] 11.572 77.2478 3.257006e-002
[22:40:33] 11.576 77.2217 3.257006e-002
[22:40:59] 11.553 77.3738 3.257006e-002
[22:41:25] 11.571 77.2501 3.257006e-002
[22:41:51] 11.567 77.2821 3.257006e-002
[22:42:18] 11.584 77.1680 3.257006e-002
[22:42:45] 11.605 77.0264 3.257006e-002
[22:43:11] 11.567 77.2788 3.257006e-002
[22:43:38] 11.585 77.1592 3.257006e-002
[22:44:04] 11.582 77.1819 3.257006e-002
[22:44:30] 11.572 77.2464 3.257006e-002
[22:44:59] 11.603 77.0389 3.257006e-002
[22:45:35] 11.609 76.9979 3.257006e-002
[22:46:12] 11.592 77.1097 3.257006e-002
[22:46:51] 11.630 76.8585 3.257006e-002
[22:47:37] 11.589 77.1293 3.257006e-002
[22:48:10] 11.797 75.7749 3.257006e-002
[22:48:44] 11.579 77.1990 3.257006e-002
Testing ended on 10/07/2012 10:48:44 PM
Test Result: Success.

CineBench R11.5

CineBench R10

WPrime 32M and 1024M

7Zip

3DMark11

3DMark Vantage

Crysis Warhead

Stalker: Call Of Pripyat

1920x1080

Unigine 3.0

POV-Ray 3.7 RC6

 

[Sweepr]

Awaiting the results.

 

[BTRY B 529th FA BN]

subscribed!

 

[jhu]

I'd like to request povray, which is currently at 3.7 RC6

 

[Makaveli]

sub'd also!

 

[alexruiz]

Would it be too much to ask for the gaming benchmarks to run at in game settings? Yes, I mean 1920 x 1080 with everything maxed out. For the 800 x 600 numbers, I can see anandtech

 

[bmadd89]

Will update pictures with max settings. Will update first post with info on benchmark
settings for other people to replicate.

 

[Termie]

Looking forward to your results!

 

[bmadd89]

AMD board and cpu out. Adjusting waterblock for intel mount and running new hoses. Getting the 3930k in a few hours so i should have that up and running before Saturday.

 

[Rhoxed]

looking forward to results

 

[bmadd89]

Started running tests and results were all over the place. Figured it out the even though windows power management was set to high performance and processor state should be 100% all the time EIST is kicking in and taking it pack down to 1900mhz. Will run benches tomorrow night as i have had 2 late nights in a row and tomorrow will be a long sesion now.

Got the 3930k and the mobo. The board has been used before (clearly) and sata cables have been used already so i have to take it back in and get a new one.

24hrs from now you should have your results. from the brief testing i have done the 980 @ 4186 certainly beats the 1090t. I will put 6 and 12 thread results up.

 

[bmadd89]

Results for the Intel i7-980 (speeds in first post.) 6cores. 6 threads. This should give the AMD as equal a chance to match the i7-980 as it can.

IBT
----------------------------
IntelBurnTest v2.53
Created by AgentGOD
----------------------------
Processor: Intel(R) Core(TM) i7 CPU 980 @ 3.33GHz
Clock Speed: 4.03 GHz
Active Physical Cores: 6
Total System Memory: 12279 MB
Stress Level: Standard (1024 MB)
Testing started on 13/07/2012 4:31:25 PM
Time (s) Speed (GFlops) Result
[16:31:41] 10.335 86.4890 3.522482e-002
[16:31:56] 10.334 86.4971 3.522482e-002
[16:32:11] 10.431 85.6976 3.522482e-002
[16:32:27] 10.337 86.4702 3.522482e-002
[16:32:42] 10.328 86.5477 3.522482e-002
[16:32:57] 10.330 86.5325 3.522482e-002
[16:33:13] 10.527 84.9131 3.522482e-002
[16:33:28] 10.350 86.3623 3.522482e-002
[16:33:44] 10.413 85.8424 3.522482e-002
[16:33:59] 10.393 86.0087 3.522482e-002
[16:34:15] 10.396 85.9849 3.522482e-002
[16:34:30] 10.299 86.7895 3.522482e-002
[16:34:45] 10.356 86.3169 3.522482e-002
[16:35:01] 10.324 86.5822 3.522482e-002
[16:35:16] 10.352 86.3481 3.522482e-002
[16:35:31] 10.348 86.3855 3.522482e-002
[16:35:47] 10.336 86.4829 3.522482e-002
[16:36:02] 10.357 86.3031 3.522482e-002
[16:36:17] 10.340 86.4529 3.522482e-002
[16:36:33] 10.311 86.6940 3.522482e-002
Testing ended on 13/07/2012 4:36:33 PM
Test Result: Success.
----------------------------

CineBench R11.5

CineBench R10

Wprime
I can't get it to complete on the 980 regardless of clock speed or core/thread count. Will keep try and update if something changes.

7Zip

3dMark11

3dMark Vantage

Crysis Warhead

Stalker: Call Of Pripyat 1080p

Unigine 3.0 1080p

POV-Ray 3.7 RC6

 

[bmadd89]

Results for the Intel i7-980 (speeds in first post.) 6cores. 12 threads. Same speed as above.

IBT
IBT kills the Intel processor once you put hyperthreading on. Sitting around 66GFlops.

CineBench R11.5

CineBench R10

7Zip

3dMark11

3dMark Vantage

Crysis Warhead 1080p and 1200p

Stalker: COP 1080p

Unigine 3.0 1080p

POV-Ray 3.7 RC6

 

[bmadd89]

7Zip

POV-Ray

CineBench R10

CineBench R11.5

 

[Termie]

The improvement in all the gaming tests, including the theoretically gpu-limited unigine and 3dmark11 graphics tests 1-4 are astounding.

 

[p_monks33]

1090T still looks like a great performer. I hear people saying if you have an AMD processor you need to upgrade. This looks to me like the 1090T is still an awesome gaming platform.

 

[Termie]

Do you find that IBT gives consistent results from test to test? Is that a built-in benchmark you're using? I haven't seen that type of output from IBT before. In stress-testing, I've found the results to vary even when nothing has changed.

 

[inf64]

1090T still looks like a great performer. I hear people saying if you have an AMD processor you need to upgrade. This looks to me like the 1090T is still an awesome gaming platform.

That's indeed true. Especially when you factor in the price.

 

[bmadd89]

IBT is "Fairly" consistent. I just run IBT in this situation to show that they were stable enough to run IBT for 20 runs (i know that doesnt prove much but it's something). The GFlops on go up or down but 1.5Glops or so at the very most. I was also had a browser window open while the IBTwas running on both systems to that would probably make it move around a but

The 1090T might be a great performer at that speed but lets be honest few other poeple will get one to that speed. I could get it a bit more but i have been using only AMD systems for nearly a decade and i had that system tweaked. 4.2Ghz with a 3Ghz NB clock is the best that it could get, on water, with 1.54V and 1.38 on the NB.

I'm still fairly new to the intel side of it but the 980 seems to have just as much legs in it but im limited by my knowledge on the platform. Given a month or so of playing around and tweaking the uncore and QPI im sure i can get a bigger gap going in the cpu bound areas.

In this graph even Stalker: COP is cpu limit by the 980 @ 4.2Ghz.

I loved the AMD systems i've had. While they don't perform top notch performance they do alright which is what i was happy with and didnt mind lossing 10% or so frames but Bulldozer is worse (hopefully piledriver is equal or better the Phenom II at the same speed) then my 1090T and Sandy and Ivy only increase the gap they had on my 1090T. In CPU limit situations they (ppl with Sandy of Ivy) could be gettting 25% more frames at the same speed from the info i have got from comparing the 1090T to the 980.

While before i would have agreed with you that people who have been telling people to upgrade to an Intel are fanboys or trolling but know i would have to back them up if push come to shove.

The bottom line is If you have a high end card or crossfire an AMD cpu regardless of speed will bottleneck you at 1920x1200 or less.

 

[moonbogg]

The bottom line is If you have a high end card or crossfire an AMD cpu regardless of speed will bottleneck you at 1920x1200 or less.

BOOYAH! Straight from the alpaca's mouth (not calling you an alpaca, its just a commonly used figure of speech). I am hurt that you didn't mention SLI setups. I am a sensitive fanboy with undying brand loyalty and the lack of SLI being mentioned has made me feel less secure with myself. I'll love you forever for seeing the light that an Intel CPU provides though. aMd = serious bottlethroating.

 

[bmadd89]

Lets be serious though. The 1090T served me well. It just gave me at best 70% what a modern Intel could give regardless how much voltage and power I push into it. AMD has a place in the world, but if you seriously care about performance you can't, honestly, with a straight face say AMD is better choice then Intel. its taken me a many years but the tipping point was the original Phenom 9500 I bought but here I am now a better person for crossing this bridge.

 

[SlowSpyder]

Lets be serious though. The 1090T served me well. It just gave me at best 70% what a modern Intel could give regardless how much voltage and power I push into it. AMD has a place in the world, but if you seriously care about performance you can't, honestly, with a straight face say AMD is better choice then Intel. its taken me a many years but the tipping point was the original Phenom 9500 I bought but here I am now a better person for crossing this bridge.

I don't think anyone says AMD is better these days. I do think a lot of people suggest that Phenom II is/was 'good enough' for most people. With that being said, if you are building a new system, it is pretty difficult to recommend AMD these days, unless there is some kind of crazy sale or something.

 

[jhu]

AMD 1090T. Speed in first post

Results for the Intel i7-980 (speeds in first post.) 6cores. 6 threads.

Thanks for putting up povray. It's amazing what recompiling can do (or how unoptimized the stock binary is). Here's what I get for my 1090T @ stock speeds (3.2 GHz, FreeBSD 9.0, 64-bit, gcc 4.6):

Render Time:
Photon Time: 0 hours 0 minutes 2 seconds (2.390 seconds)
using 9 thread(s) with 2.763 CPU-seconds total
Radiosity Time: No radiosity
Trace Time: 0 hours 3 minutes 10 seconds (190.466 seconds)
using 6 thread(s) with 1113.568 CPU-seconds total

Which ends up to be 1376.3 pps