Share your AIDA64 GPGPU Benchmark Results

-rascal- · April 12, 2016

Hey everyone,

I am looking into GPU computing related comparisons for my personal academic purposes (Electrical Engineering), and am wondering if anyone is willing to share their numbers.

Basically, I have been digging into distributed computing, real-time machine learning, Neural Networks, and computing cloud services (i.e. Amazon AWS, and Google GCE) -- and if it is possible to make a dedicated FPGA to compete with GPUs, but with a lower power consumption and heat output. I mean, heck, think of four GTX Titan X's in one system....and then multiply that by four systems. The circuit breaker for my house will kill itself. If anybody is curious:

http://pjreddie.com/darknet/hardware-guide/

Please, if anyone is willing to spend 3~5 minutes, and run AIDA64's built-in GPGPU Benchmark, and post their results. I understand this won't give "bananas-to-bananas" values, due to so many system variables, but it will give me a "quick-and-dirty" comparison. It is better than for me to spend money on one of every AMD or NVIDIA GPU just to get some quick numbers -- I am not like ol' Linus with containers of GPUs...though I might have to do so if I eventually get serious, and fully commit to this.

EDIT #1: Please, have a clean environment, if possible, so the GPU can dedicate all/most of its resources, cores, and priorities to the benchmark.

EDIT #2: The .txt doesn't indicate WHAT GPU you are using, (and the frequencies) so I will need to know that. Thank you for the results so far, you beautiful people!

A screenshot of the results will be fine, but it would be amazing if you could include the .txt results file as well.

For an example, with my R9-Fury (see attached files)

Thank you!!

gpgpu_1.txt

Moonzy · April 12, 2016

`-`

gpgpu i5 6600 + 970.txt

Manswaggle · April 12, 2016

Same video card as in my signature, 290, but it's clocked to 1055/1400 right now(troubleshooting and too lazy to clock it back).

Edit: Image was taking up a lot of space.

gpgpu.txt

Nacho Marco Segui · April 12, 2016

Ayy.txt

-rascal- · April 12, 2016

10 minutes ago, Moonzy said:

`-`

[-- Sexy Pic --]

gpgpu i5 6600 + 970.txt

Thanks bruh!

Some of your GTX 970 results seems a little low for some reason. Are you running a game or something on the side?

@Nacho Marco Segui is this your GTX 970 @ 1550/8000?

Moonzy · April 12, 2016

1 minute ago, -rascal- said:

Thanks bruh!

Some of your GTX 970 results seems a little low for some reason. Are you running a game or something on the side?

i dont mind giving u a result from clean environment

no games running but i have a couple paused videos `-`

need a clean environment result??

Moonzy · April 12, 2016

2 minutes ago, -rascal- said:

Thanks bruh!

Some of your GTX 970 results seems a little low for some reason. Are you running a game or something on the side?

and overclocked or stock `-`

(1316 vs 1540)

oskarha · April 12, 2016

I will try to remember to upload some numbers from my already benchmarked machines when i get home later today.

Already benchmarked results:

GTX 660

Quadro 2000

Radeon HD 6770m

GT 430

And if I get time to through these cards on my testbench:

GTX 280

Radeon HD 6950 (unlocked to 6970)

Radeon HD 5750

Tell me if your interested in these super low end gpus:

NVS 300

Radeon HD 5450

Radeon HD 3470

Radeon HD 3450

Radeon HD 2400 XT

-rascal- · April 12, 2016

1 minute ago, Moonzy said:

i dont mind giving u a result from clean environment

no games running but i have a couple paused videos `-`

need a clean environment result??

Right, I should mention that in the original post just in case.

Seems like it is best to have a clean environment.

I don't care about the clock speeds at this stage.

It's hard to gauge GPUs now, especially with factory overclocks, and automatic Boost clocks, etc.

Manswaggle · April 12, 2016

I increased the clocks, idk if the variable will interest you. Went from 1055 on the core to 1120 and 1400 to 1450 on memory.

Edit: Did a clean run.

gpgpu1.txt

gpgpu_clean.txt

-rascal- · April 12, 2016

Just now, Manswaggle said:

I increased the clocks, idk if the variable will interest you. Went from 1055 on the core to 1120 and 1400 to 1450 on memory.

gpgpu1.txt

All good, all good.

Thank you .

JPotze · April 12, 2016

i have a trial version so dont know if that impacts anything. gtx 970 watercooled, stock speed.

gpgpu.txt

-rascal- · April 12, 2016

2 minutes ago, JPotze said:

i have a trial version so dont know if that impacts anything. gtx 970 watercooled, stock speed.

gpgpu.txt

Those numbers look okay -- at least compared to the one and only GTX 970 result I can find...

Moonzy · April 12, 2016

i5 6600 + GTX 970 Asus Strix (1540 core clock/7650 mem clock)

with minimal things running (avast and microsoft's own stuffs `-`)

gpgpu i5 6600 + 970.txt

-rascal- · April 12, 2016

2 minutes ago, Moonzy said:

i5 6600 + GTX 970 Asus Strix (1540 core clock/7650 mem clock)

with minimal things running (avast and microsoft's own stuffs `-`)

gpgpu i5 6600 + 970.txt

There it is!

You and @JPotze's GTX 970 result looks good -- of course, your overclocked GPU pulls a bit ahead in the numbers

Nacho Marco Segui · April 12, 2016

35 minutes ago, -rascal- said:

Thanks bruh!

Some of your GTX 970 results seems a little low for some reason. Are you running a game or something on the side?

@Nacho Marco Segui is this your GTX 970 @ 1550/8000?

Only +200 on the memory, running at 1404 on the core

Moonzy · April 12, 2016

here's my result on stock `-`

GTX 970 (1316/7010)

gpgpu i5 6600 + 970 stock.txt

are you by any chance interested in mobile gpus as well?

Manswaggle · April 12, 2016

I was going to try to test it on a GTX 460, but AIDA didn't want to.

xAcid9 · April 12, 2016

R9-290x + Xeon E3-1230v2

Stock

Spoiler

Mild OC

Spoiler

Benchmark                        Result  Run Time  Build Time
-------------------------------------------------------------
Memory Read                  11424 MB/s   1296 ms            
- Pinned                     11424 MB/s      3 ms            
- Pageable                    6433 MB/s      5 ms            
Memory Write                 11380 MB/s   1328 ms            
- Pinned                     11380 MB/s      3 ms            
- Pageable                    6335 MB/s      5 ms            
Memory Copy                 338222 MB/s  12969 ms            
- 15 MB Block               231848 MB/s      0 ms            
- 32 MB Block               280997 MB/s      0 ms            
- 64 MB Block               308299 MB/s      0 ms            
- 128 MB Block              309339 MB/s      1 ms            
- 256 MB Block              327709 MB/s      2 ms            
- 512 MB Block              337737 MB/s      3 ms            
- 1024 MB Block             338222 MB/s      6 ms            
Single-Precision FLOPS      6070 GFLOPS   8203 ms            
- float1                    6070 GFLOPS    725 ms      125 ms
- float2                    6070 GFLOPS    725 ms      125 ms
- float4                    5942 GFLOPS    740 ms      125 ms
- float8                    5942 GFLOPS    740 ms      125 ms
- float16                   5937 GFLOPS    741 ms      141 ms
Double-Precision FLOPS     758.8 GFLOPS   8297 ms            
- double1                  758.8 GFLOPS    724 ms      172 ms
- double2                  743.0 GFLOPS    740 ms      140 ms
- double4                  742.9 GFLOPS    740 ms      140 ms
- double8                  742.9 GFLOPS    740 ms      125 ms
- double16                 742.8 GFLOPS    740 ms      172 ms
24-bit Integer IOPS          6066 GIOPS   7828 ms            
- int1                       6066 GIOPS    725 ms       31 ms
- int2                       6066 GIOPS    725 ms       47 ms
- int4                       5939 GIOPS    741 ms       63 ms
- int8                       5940 GIOPS    740 ms       78 ms
- int16                      5938 GIOPS    741 ms      141 ms
32-bit Integer IOPS          2021 GIOPS   8000 ms            
- int1                       2021 GIOPS    544 ms     1359 ms
- int2                       2019 GIOPS    545 ms      156 ms
- int4                       1974 GIOPS    557 ms      172 ms
- int8                       1968 GIOPS    559 ms      234 ms
- int16                      1957 GIOPS    562 ms      390 ms
64-bit Integer IOPS         505.6 GIOPS   8578 ms            
- long1                     505.6 GIOPS    544 ms     1312 ms
- long2                     493.6 GIOPS    557 ms      157 ms
- long4                     492.2 GIOPS    558 ms      234 ms
- long8                     424.0 GIOPS    648 ms      375 ms
- long16                    274.4 GIOPS    501 ms      765 ms
AES-256                      32579 MB/s   9469 ms            
- inline loop / 4 MB         31437 MB/s    521 ms       78 ms
- inline loop / 8 MB         32189 MB/s    509 ms            
- inline loop / 16 MB        32231 MB/s    508 ms            
- inline loop / 32 MB        32427 MB/s    505 ms            
- inline loop / 64 MB        32527 MB/s    504 ms            
- inline loop / 128 MB       32579 MB/s    503 ms            
- inline loop / 256 MB       32577 MB/s    503 ms            
SHA-1 Hash                   73684 MB/s  15016 ms            
- 512 blocks                  3274 MB/s    611 ms      610 ms
- 1024 blocks                 6528 MB/s    613 ms            
- 2048 blocks                13068 MB/s    612 ms            
- 4096 blocks                26095 MB/s    613 ms            
- 8192 blocks                52136 MB/s    614 ms            
- 16384 blocks               73684 MB/s    869 ms            
- 32768 blocks               12185 MB/s    657 ms            
- 65536 blocks               10462 MB/s    765 ms            
- 131072 blocks              10578 MB/s    756 ms            
Single-Precision Julia        933.1 FPS   9765 ms            
- float1 break                756.3 FPS    677 ms       63 ms
- float1 stay / unroll 3      717.8 FPS    713 ms       31 ms
- float2 stay / unroll 3      924.2 FPS    554 ms       47 ms
- float4 stay / unroll 3      765.6 FPS    669 ms       46 ms
- float1 stay / unroll 9      714.0 FPS    717 ms       47 ms
- float2 stay / unroll 9      933.1 FPS    549 ms       47 ms
- float4 stay / unroll 9      771.2 FPS    664 ms       63 ms
Double-Precision Mandel       103.5 FPS  12219 ms            
- double1 break                84.2 FPS    760 ms       47 ms
- double1 stay / unroll 3      83.3 FPS    768 ms       47 ms
- double2 stay / unroll 3     102.0 FPS    628 ms       31 ms
- double4 stay / unroll 3     103.5 FPS    618 ms       47 ms
- double8 stay / unroll 3      70.2 FPS    912 ms       78 ms
- double1 stay / unroll 9      84.1 FPS    761 ms       47 ms
- double2 stay / unroll 9     100.0 FPS    640 ms       47 ms
- double4 stay / unroll 9      98.0 FPS    653 ms       79 ms

-rascal- · April 12, 2016

8 hours ago, Manswaggle said:

I was going to try to test it on a GTX 460, but AIDA didn't want to.

What?! Why?

A GTX 460 should be compatible...

4 hours ago, xAcid9 said:

R9-290x + Xeon E3-1230v2

Stock

Reveal hidden contents

Mild OC

Reveal hidden contents


Benchmark                        Result  Run Time  Build Time
-------------------------------------------------------------
Memory Read                  11424 MB/s   1296 ms            
- Pinned                     11424 MB/s      3 ms            
- Pageable                    6433 MB/s      5 ms            
Memory Write                 11380 MB/s   1328 ms            
- Pinned                     11380 MB/s      3 ms            
- Pageable                    6335 MB/s      5 ms            
Memory Copy                 338222 MB/s  12969 ms            
- 15 MB Block               231848 MB/s      0 ms            
- 32 MB Block               280997 MB/s      0 ms            
- 64 MB Block               308299 MB/s      0 ms            
- 128 MB Block              309339 MB/s      1 ms            
- 256 MB Block              327709 MB/s      2 ms            
- 512 MB Block              337737 MB/s      3 ms            
- 1024 MB Block             338222 MB/s      6 ms            
Single-Precision FLOPS      6070 GFLOPS   8203 ms            
- float1                    6070 GFLOPS    725 ms      125 ms
- float2                    6070 GFLOPS    725 ms      125 ms
- float4                    5942 GFLOPS    740 ms      125 ms
- float8                    5942 GFLOPS    740 ms      125 ms
- float16                   5937 GFLOPS    741 ms      141 ms
Double-Precision FLOPS     758.8 GFLOPS   8297 ms            
- double1                  758.8 GFLOPS    724 ms      172 ms
- double2                  743.0 GFLOPS    740 ms      140 ms
- double4                  742.9 GFLOPS    740 ms      140 ms
- double8                  742.9 GFLOPS    740 ms      125 ms
- double16                 742.8 GFLOPS    740 ms      172 ms
24-bit Integer IOPS          6066 GIOPS   7828 ms            
- int1                       6066 GIOPS    725 ms       31 ms
- int2                       6066 GIOPS    725 ms       47 ms
- int4                       5939 GIOPS    741 ms       63 ms
- int8                       5940 GIOPS    740 ms       78 ms
- int16                      5938 GIOPS    741 ms      141 ms
32-bit Integer IOPS          2021 GIOPS   8000 ms            
- int1                       2021 GIOPS    544 ms     1359 ms
- int2                       2019 GIOPS    545 ms      156 ms
- int4                       1974 GIOPS    557 ms      172 ms
- int8                       1968 GIOPS    559 ms      234 ms
- int16                      1957 GIOPS    562 ms      390 ms
64-bit Integer IOPS         505.6 GIOPS   8578 ms            
- long1                     505.6 GIOPS    544 ms     1312 ms
- long2                     493.6 GIOPS    557 ms      157 ms
- long4                     492.2 GIOPS    558 ms      234 ms
- long8                     424.0 GIOPS    648 ms      375 ms
- long16                    274.4 GIOPS    501 ms      765 ms
AES-256                      32579 MB/s   9469 ms            
- inline loop / 4 MB         31437 MB/s    521 ms       78 ms
- inline loop / 8 MB         32189 MB/s    509 ms            
- inline loop / 16 MB        32231 MB/s    508 ms            
- inline loop / 32 MB        32427 MB/s    505 ms            
- inline loop / 64 MB        32527 MB/s    504 ms            
- inline loop / 128 MB       32579 MB/s    503 ms            
- inline loop / 256 MB       32577 MB/s    503 ms            
SHA-1 Hash                   73684 MB/s  15016 ms            
- 512 blocks                  3274 MB/s    611 ms      610 ms
- 1024 blocks                 6528 MB/s    613 ms            
- 2048 blocks                13068 MB/s    612 ms            
- 4096 blocks                26095 MB/s    613 ms            
- 8192 blocks                52136 MB/s    614 ms            
- 16384 blocks               73684 MB/s    869 ms            
- 32768 blocks               12185 MB/s    657 ms            
- 65536 blocks               10462 MB/s    765 ms            
- 131072 blocks              10578 MB/s    756 ms            
Single-Precision Julia        933.1 FPS   9765 ms            
- float1 break                756.3 FPS    677 ms       63 ms
- float1 stay / unroll 3      717.8 FPS    713 ms       31 ms
- float2 stay / unroll 3      924.2 FPS    554 ms       47 ms
- float4 stay / unroll 3      765.6 FPS    669 ms       46 ms
- float1 stay / unroll 9      714.0 FPS    717 ms       47 ms
- float2 stay / unroll 9      933.1 FPS    549 ms       47 ms
- float4 stay / unroll 9      771.2 FPS    664 ms       63 ms
Double-Precision Mandel       103.5 FPS  12219 ms            
- double1 break                84.2 FPS    760 ms       47 ms
- double1 stay / unroll 3      83.3 FPS    768 ms       47 ms
- double2 stay / unroll 3     102.0 FPS    628 ms       31 ms
- double4 stay / unroll 3     103.5 FPS    618 ms       47 ms
- double8 stay / unroll 3      70.2 FPS    912 ms       78 ms
- double1 stay / unroll 9      84.1 FPS    761 ms       47 ms
- double2 stay / unroll 9     100.0 FPS    640 ms       47 ms
- double4 stay / unroll 9      98.0 FPS    653 ms       79 ms

Awesome! Thank you!

Manswaggle · April 12, 2016

@-rascal-It said something along the lines of no Open_CL compatible card. Which, I don't really get why, 460 should be compatible with Open_CL.

-rascal- · April 13, 2016

Thank you for everybody who is willing to share their results so far!

Please keep them coming.

P.S.

I don't mind if we could get some numbers for the GTX 980 / 980Ti, Titan X, or even GTX 780 / 780Ti.

Of course, a few more R9-200 or R9-300 series GPUs, and probably throw in a few R9-Fury X / Fury / Nano.

Once again, thank you!

Sign In

Share your AIDA64 GPGPU Benchmark Results

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites

Link to post

Share on other sites

Link to comment

Share on other sites