Perpetual benchmark thread... - Page 49

abstractalgebra · 2011-01-20, 07:03

I5-2500k Sandy Bridge Benchmark
Turbo Overclocked to 4.1GHz, 1333 DDR3, Unbuntu 10.10, mprime 2511-64bit.

Stock cooler, will post more benchmarks speed increases. 4.4-4.6 appears to be common with ~5Ghz if rather lucky

Hit enter to continue: [Main thread Jan 20 01:52] Starting worker.
[Worker #1 Jan 20 01:52] Worker starting
[Worker #1 Jan 20 01:52] Setting affinity to run worker on logical CPU #0
[Worker #1 Jan 20 01:52] Your timings will be written to the results.txt file.
[Worker #1 Jan 20 01:52] Compare your results to other computers at http://www.mersenne.org/report_benchmarks
[Worker #1 Jan 20 01:52] Timing 55 iterations at 768K FFT length. Best time: 6.818 ms.
[Worker #1 Jan 20 01:52] Timing 47 iterations at 896K FFT length. Best time: 8.416 ms.
[Worker #1 Jan 20 01:52] Timing 41 iterations at 1024K FFT length. Best time: 9.397 ms.
[Worker #1 Jan 20 01:53] Timing 33 iterations at 1280K FFT length. Best time: 11.923 ms.
[Worker #1 Jan 20 01:53] Timing 27 iterations at 1536K FFT length. Best time: 14.280 ms.
[Worker #1 Jan 20 01:53] Timing 23 iterations at 1792K FFT length. Best time: 17.603 ms.
[Worker #1 Jan 20 01:53] Timing 20 iterations at 2048K FFT length. Best time: 19.479 ms.
[Worker #1 Jan 20 01:53] Timing 16 iterations at 2560K FFT length. Best time: 26.130 ms.
[Worker #1 Jan 20 01:53] Timing 13 iterations at 3072K FFT length. Best time: 31.586 ms.
[Worker #1 Jan 20 01:53] Timing 11 iterations at 3584K FFT length. Best time: 38.991 ms.
[Worker #1 Jan 20 01:53] Timing 10 iterations at 4096K FFT length. Best time: 43.271 ms.
[Worker #1 Jan 20 01:53] Timing 10 iterations at 5120K FFT length. Best time: 55.765 ms.
[Worker #1 Jan 20 01:53] Timing 10 iterations at 6144K FFT length. Best time: 66.860 ms.
[Worker #1 Jan 20 01:53] Timing 10 iterations at 7168K FFT length. Best time: 81.550 ms.
[Worker #1 Jan 20 01:53] Timing 10 iterations at 8192K FFT length. Best time: 91.063 ms.

willmore · 2011-01-21, 15:42

Quote:

Originally Posted by Colt45ws

The way I interpret this, is P95 is utilizing the capability of an i7 core well enough, that there is not much in the way of resources left for the 'fake' thread to utilize, or the way its splitting the load does not work well with it.

I'm curious of this holds true for SB. I'm looking to build out an SB system in the near future and I'm not sure the extra $100 for a 2600K vs a 2500K is worth it just for HT. I'm going to move this into my existing water cooled box, so I imagine they'll both overclock just as well, so all that leaves is HT or no HT.

Anyone got an opinion on this?

Prime95 · 2011-01-21, 17:15

Quote:

Originally Posted by willmore

Anyone got an opinion on this?

For prime95 I would not get HT. If you do other projects like NFS sieving or GMP-ECM, the HT might be worthwhile.

mdettweiler · 2011-01-21, 22:20

Quote:

Originally Posted by Prime95

For prime95 I would not get HT. If you do other projects like NFS sieving or GMP-ECM, the HT might be worthwhile.

Just curious...does this conclusion also apply for smaller FFT work as done by LLR? A friend of mine on the forum has an i7 and says that he gets ~5 cores' worth of speed on LLR because of the HT, but some of that could just be because his non-HT reference for comparison is a somewhat older Q6600.

abstractalgebra · 2011-01-22, 15:43

i5-2500K overclocked by turbo to 4.59Ghz w/1600Mhz XMP RAM (defaults)

[Sat Jan 22 10:38:54 2011]
Compare your results to other computers at http://www.mersenne.org/report_benchmarks
Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU speed: 3292.69 MHz, 4 cores
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2, SSE4
L1 cache size: unknown
L2 cache size: 256 KB, L3 cache size: 6 MB
L1 cache line size: unknown
L2 cache line size: 64 bytes
TLBS: 64
Prime95 64-bit version 26.4, RdtscTiming=1
Best time for 768K FFT length: 5.598 ms., avg: 5.699 ms.
Best time for 896K FFT length: 6.896 ms., avg: 6.935 ms.
Best time for 1024K FFT length: 7.514 ms., avg: 7.639 ms.
Best time for 1280K FFT length: 9.756 ms., avg: 9.911 ms.
Best time for 1536K FFT length: 11.813 ms., avg: 11.993 ms.
Best time for 1792K FFT length: 14.354 ms., avg: 14.370 ms.
Best time for 2048K FFT length: 15.807 ms., avg: 15.844 ms.
Best time for 2560K FFT length: 20.169 ms., avg: 20.199 ms.
Best time for 3072K FFT length: 24.739 ms., avg: 24.791 ms.
Best time for 3584K FFT length: 29.892 ms., avg: 29.951 ms.
Best time for 4096K FFT length: 33.029 ms., avg: 33.077 ms.
Best time for 5120K FFT length: 42.875 ms., avg: 42.908 ms.
Best time for 6144K FFT length: 53.270 ms., avg: 53.807 ms.
Best time for 7168K FFT length: 65.453 ms., avg: 65.904 ms.
Best time for 8192K FFT length: 71.324 ms., avg: 72.243 ms.
Timing FFTs using 2 threads.
Best time for 768K FFT length: 3.022 ms., avg: 3.048 ms.
Best time for 896K FFT length: 3.677 ms., avg: 3.856 ms.
Best time for 1024K FFT length: 4.035 ms., avg: 4.128 ms.
Best time for 1280K FFT length: 5.150 ms., avg: 5.292 ms.
Best time for 1536K FFT length: 6.131 ms., avg: 6.144 ms.
Best time for 1792K FFT length: 7.581 ms., avg: 7.793 ms.
Best time for 2048K FFT length: 8.344 ms., avg: 8.379 ms.
Best time for 2560K FFT length: 10.628 ms., avg: 10.903 ms.
Best time for 3072K FFT length: 12.972 ms., avg: 13.322 ms.
Best time for 3584K FFT length: 15.672 ms., avg: 15.881 ms.
Best time for 4096K FFT length: 17.281 ms., avg: 17.750 ms.
Best time for 5120K FFT length: 22.446 ms., avg: 22.583 ms.
Best time for 6144K FFT length: 27.890 ms., avg: 28.191 ms.
Best time for 7168K FFT length: 34.689 ms., avg: 35.205 ms.
Best time for 8192K FFT length: 37.308 ms., avg: 37.414 ms.
Timing FFTs using 3 threads.
Best time for 768K FFT length: 2.078 ms., avg: 2.122 ms.
Best time for 896K FFT length: 2.563 ms., avg: 2.664 ms.
Best time for 1024K FFT length: 2.816 ms., avg: 2.968 ms.
Best time for 1280K FFT length: 3.561 ms., avg: 3.649 ms.
Best time for 1536K FFT length: 4.219 ms., avg: 4.327 ms.
Best time for 1792K FFT length: 5.210 ms., avg: 5.251 ms.
Best time for 2048K FFT length: 5.765 ms., avg: 5.801 ms.
Best time for 2560K FFT length: 7.328 ms., avg: 7.624 ms.
Best time for 3072K FFT length: 8.937 ms., avg: 9.280 ms.
Best time for 3584K FFT length: 10.750 ms., avg: 10.768 ms.
Best time for 4096K FFT length: 11.897 ms., avg: 12.342 ms.
Best time for 5120K FFT length: 15.377 ms., avg: 15.549 ms.
Best time for 6144K FFT length: 18.939 ms., avg: 19.378 ms.
Best time for 7168K FFT length: 25.291 ms., avg: 26.186 ms.
Best time for 8192K FFT length: 25.706 ms., avg: 25.900 ms.
Timing FFTs using 4 threads.
Best time for 768K FFT length: 1.738 ms., avg: 1.868 ms.
Best time for 896K FFT length: 1.986 ms., avg: 2.019 ms.
Best time for 1024K FFT length: 2.238 ms., avg: 2.482 ms.
Best time for 1280K FFT length: 2.781 ms., avg: 2.986 ms.
Best time for 1536K FFT length: 3.345 ms., avg: 3.837 ms.
Best time for 1792K FFT length: 4.082 ms., avg: 4.319 ms.
Best time for 2048K FFT length: 4.515 ms., avg: 4.792 ms.
Best time for 2560K FFT length: 5.774 ms., avg: 6.054 ms.
Best time for 3072K FFT length: 6.942 ms., avg: 7.309 ms.
Best time for 3584K FFT length: 8.422 ms., avg: 8.582 ms.
Best time for 4096K FFT length: 9.295 ms., avg: 9.801 ms.
Best time for 5120K FFT length: 11.995 ms., avg: 12.454 ms.
Best time for 6144K FFT length: 14.655 ms., avg: 14.902 ms.
Best time for 7168K FFT length: 23.160 ms., avg: 23.649 ms.
Best time for 8192K FFT length: 19.803 ms., avg: 20.997 ms.
Best time for 58 bit trial factors: 1.453 ms.
Best time for 59 bit trial factors: 1.504 ms.
Best time for 60 bit trial factors: 1.502 ms.
Best time for 61 bit trial factors: 1.687 ms.
Best time for 62 bit trial factors: 1.675 ms.
Best time for 63 bit trial factors: 1.896 ms.
Best time for 64 bit trial factors: 1.983 ms.
Best time for 65 bit trial factors: 2.307 ms.
Best time for 66 bit trial factors: 2.729 ms.
Best time for 67 bit trial factors: 2.695 ms.

http://valid.canardpc.com/show_oc.php?id=1611553

James Heinrich · 2011-01-22, 18:49

Quote:

Originally Posted by abstractalgebra

i5-2500K overclocked by turbo to 4.59Ghz w/1600Mhz XMP RAM (defaults)
...
Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU speed: 3292.69 MHz, 4 cores

So which is it? 3.3GHz or 4.6GHz?

abstractalgebra · 2011-01-22, 21:48

Quote:

Originally Posted by James Heinrich

So which is it? 3.3GHz or 4.6GHz?

Fair question, the CPU defaults to 3.3GHz but my overlocked turbo boosts to 4.6 GHz. With Sandy Bridge you can't just increase the default GHz, so it would be nice for Prime95 to implement a CPU-Z style check of the current speed. (see the validated CPU speed link in my other post)

To make matters more complex, you can set different turbo boost multipliers for 1,2,3,4 cores being active. In my benchmark I set all cores at the same 46x = 4.6GHz. However, for example, if only one core was active I could setup a 48x multiplier to get one core at 4.8Ghz.

Commaster · 2011-01-23, 01:14

By the way, which is more stable: a higher Base Clock or an insanely huge Multiplier? (Speaking only about the CPU itself)

James Heinrich · 2011-01-23, 01:32

Quote:

Originally Posted by Commaster

By the way, which is more stable: a higher Base Clock or an insanely huge Multiplier? (Speaking only about the CPU itself)

Isn't that kind of a non-issue now with SandyBridge? You're basically now limited (as I understand it) to playing with the multiplier; adjusting the base clock by anything other than a tiny amount is no longer feasible since it affects everything, including things like PCIe that don't like to be overclocked.

lavalamp · 2011-01-23, 03:50

Quote:

Originally Posted by abstractalgebra

Fair question, the CPU defaults to 3.3GHz but my overlocked turbo boosts to 4.6 GHz. With Sandy Bridge you can't just increase the default GHz, so it would be nice for Prime95 to implement a CPU-Z style check of the current speed.

I don't know if it is the same for all boards, but on Gigabyte boards at least, you can set the max turbo mode CPU multiplier for 1 - 4 cores, or you can disable turbo mode entirely and set a constant multiplier for all cores.

James Heinrich · 2011-01-25, 03:07

It's now possible to directly compare between CPUs on my throughput calculator. For example, assuming identical clockspeeds, you can see how an i7-2500K neatly outruns an i7-965 by a noticable amount.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Perpetual "interesting video" thread...	Xyzzy	Lounge	57	2023-06-08 04:11
LLR benchmark thread	Oddball	Riesel Prime Search	5	2010-08-02 00:11
Perpetual I'm pi**ed off thread	rogue	Soap Box	19	2009-10-28 19:17
Perpetual autostereogram thread...	Xyzzy	Lounge	10	2006-09-28 00:36
Perpetual ECM factoring challenge thread...	Xyzzy	Factoring	65	2005-09-05 08:16

2011-01-20, 07:03	#529
abstractalgebra Jan 2011 2² Posts	I5-2500k Sandy Bridge Benchmark Turbo Overclocked to 4.1GHz, 1333 DDR3, Unbuntu 10.10, mprime 2511-64bit. Stock cooler, will post more benchmarks speed increases. 4.4-4.6 appears to be common with ~5Ghz if rather lucky Hit enter to continue: [Main thread Jan 20 01:52] Starting worker. [Worker #1 Jan 20 01:52] Worker starting [Worker #1 Jan 20 01:52] Setting affinity to run worker on logical CPU #0 [Worker #1 Jan 20 01:52] Your timings will be written to the results.txt file. [Worker #1 Jan 20 01:52] Compare your results to other computers at http://www.mersenne.org/report_benchmarks [Worker #1 Jan 20 01:52] Timing 55 iterations at 768K FFT length. Best time: 6.818 ms. [Worker #1 Jan 20 01:52] Timing 47 iterations at 896K FFT length. Best time: 8.416 ms. [Worker #1 Jan 20 01:52] Timing 41 iterations at 1024K FFT length. Best time: 9.397 ms. [Worker #1 Jan 20 01:53] Timing 33 iterations at 1280K FFT length. Best time: 11.923 ms. [Worker #1 Jan 20 01:53] Timing 27 iterations at 1536K FFT length. Best time: 14.280 ms. [Worker #1 Jan 20 01:53] Timing 23 iterations at 1792K FFT length. Best time: 17.603 ms. [Worker #1 Jan 20 01:53] Timing 20 iterations at 2048K FFT length. Best time: 19.479 ms. [Worker #1 Jan 20 01:53] Timing 16 iterations at 2560K FFT length. Best time: 26.130 ms. [Worker #1 Jan 20 01:53] Timing 13 iterations at 3072K FFT length. Best time: 31.586 ms. [Worker #1 Jan 20 01:53] Timing 11 iterations at 3584K FFT length. Best time: 38.991 ms. [Worker #1 Jan 20 01:53] Timing 10 iterations at 4096K FFT length. Best time: 43.271 ms. [Worker #1 Jan 20 01:53] Timing 10 iterations at 5120K FFT length. Best time: 55.765 ms. [Worker #1 Jan 20 01:53] Timing 10 iterations at 6144K FFT length. Best time: 66.860 ms. [Worker #1 Jan 20 01:53] Timing 10 iterations at 7168K FFT length. Best time: 81.550 ms. [Worker #1 Jan 20 01:53] Timing 10 iterations at 8192K FFT length. Best time: 91.063 ms.

2011-01-22, 15:43	#533
abstractalgebra Jan 2011 2² Posts	i5-2500K overclocked by turbo to 4.59Ghz w/1600Mhz XMP RAM (defaults) [Sat Jan 22 10:38:54 2011] Compare your results to other computers at http://www.mersenne.org/report_benchmarks Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz CPU speed: 3292.69 MHz, 4 cores CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2, SSE4 L1 cache size: unknown L2 cache size: 256 KB, L3 cache size: 6 MB L1 cache line size: unknown L2 cache line size: 64 bytes TLBS: 64 Prime95 64-bit version 26.4, RdtscTiming=1 Best time for 768K FFT length: 5.598 ms., avg: 5.699 ms. Best time for 896K FFT length: 6.896 ms., avg: 6.935 ms. Best time for 1024K FFT length: 7.514 ms., avg: 7.639 ms. Best time for 1280K FFT length: 9.756 ms., avg: 9.911 ms. Best time for 1536K FFT length: 11.813 ms., avg: 11.993 ms. Best time for 1792K FFT length: 14.354 ms., avg: 14.370 ms. Best time for 2048K FFT length: 15.807 ms., avg: 15.844 ms. Best time for 2560K FFT length: 20.169 ms., avg: 20.199 ms. Best time for 3072K FFT length: 24.739 ms., avg: 24.791 ms. Best time for 3584K FFT length: 29.892 ms., avg: 29.951 ms. Best time for 4096K FFT length: 33.029 ms., avg: 33.077 ms. Best time for 5120K FFT length: 42.875 ms., avg: 42.908 ms. Best time for 6144K FFT length: 53.270 ms., avg: 53.807 ms. Best time for 7168K FFT length: 65.453 ms., avg: 65.904 ms. Best time for 8192K FFT length: 71.324 ms., avg: 72.243 ms. Timing FFTs using 2 threads. Best time for 768K FFT length: 3.022 ms., avg: 3.048 ms. Best time for 896K FFT length: 3.677 ms., avg: 3.856 ms. Best time for 1024K FFT length: 4.035 ms., avg: 4.128 ms. Best time for 1280K FFT length: 5.150 ms., avg: 5.292 ms. Best time for 1536K FFT length: 6.131 ms., avg: 6.144 ms. Best time for 1792K FFT length: 7.581 ms., avg: 7.793 ms. Best time for 2048K FFT length: 8.344 ms., avg: 8.379 ms. Best time for 2560K FFT length: 10.628 ms., avg: 10.903 ms. Best time for 3072K FFT length: 12.972 ms., avg: 13.322 ms. Best time for 3584K FFT length: 15.672 ms., avg: 15.881 ms. Best time for 4096K FFT length: 17.281 ms., avg: 17.750 ms. Best time for 5120K FFT length: 22.446 ms., avg: 22.583 ms. Best time for 6144K FFT length: 27.890 ms., avg: 28.191 ms. Best time for 7168K FFT length: 34.689 ms., avg: 35.205 ms. Best time for 8192K FFT length: 37.308 ms., avg: 37.414 ms. Timing FFTs using 3 threads. Best time for 768K FFT length: 2.078 ms., avg: 2.122 ms. Best time for 896K FFT length: 2.563 ms., avg: 2.664 ms. Best time for 1024K FFT length: 2.816 ms., avg: 2.968 ms. Best time for 1280K FFT length: 3.561 ms., avg: 3.649 ms. Best time for 1536K FFT length: 4.219 ms., avg: 4.327 ms. Best time for 1792K FFT length: 5.210 ms., avg: 5.251 ms. Best time for 2048K FFT length: 5.765 ms., avg: 5.801 ms. Best time for 2560K FFT length: 7.328 ms., avg: 7.624 ms. Best time for 3072K FFT length: 8.937 ms., avg: 9.280 ms. Best time for 3584K FFT length: 10.750 ms., avg: 10.768 ms. Best time for 4096K FFT length: 11.897 ms., avg: 12.342 ms. Best time for 5120K FFT length: 15.377 ms., avg: 15.549 ms. Best time for 6144K FFT length: 18.939 ms., avg: 19.378 ms. Best time for 7168K FFT length: 25.291 ms., avg: 26.186 ms. Best time for 8192K FFT length: 25.706 ms., avg: 25.900 ms. Timing FFTs using 4 threads. Best time for 768K FFT length: 1.738 ms., avg: 1.868 ms. Best time for 896K FFT length: 1.986 ms., avg: 2.019 ms. Best time for 1024K FFT length: 2.238 ms., avg: 2.482 ms. Best time for 1280K FFT length: 2.781 ms., avg: 2.986 ms. Best time for 1536K FFT length: 3.345 ms., avg: 3.837 ms. Best time for 1792K FFT length: 4.082 ms., avg: 4.319 ms. Best time for 2048K FFT length: 4.515 ms., avg: 4.792 ms. Best time for 2560K FFT length: 5.774 ms., avg: 6.054 ms. Best time for 3072K FFT length: 6.942 ms., avg: 7.309 ms. Best time for 3584K FFT length: 8.422 ms., avg: 8.582 ms. Best time for 4096K FFT length: 9.295 ms., avg: 9.801 ms. Best time for 5120K FFT length: 11.995 ms., avg: 12.454 ms. Best time for 6144K FFT length: 14.655 ms., avg: 14.902 ms. Best time for 7168K FFT length: 23.160 ms., avg: 23.649 ms. Best time for 8192K FFT length: 19.803 ms., avg: 20.997 ms. Best time for 58 bit trial factors: 1.453 ms. Best time for 59 bit trial factors: 1.504 ms. Best time for 60 bit trial factors: 1.502 ms. Best time for 61 bit trial factors: 1.687 ms. Best time for 62 bit trial factors: 1.675 ms. Best time for 63 bit trial factors: 1.896 ms. Best time for 64 bit trial factors: 1.983 ms. Best time for 65 bit trial factors: 2.307 ms. Best time for 66 bit trial factors: 2.729 ms. Best time for 67 bit trial factors: 2.695 ms. http://valid.canardpc.com/show_oc.php?id=1611553

2011-01-23, 01:14	#536
Commaster Jun 2010 Kiev, Ukraine 3·19 Posts	By the way, which is more stable: a higher Base Clock or an insanely huge Multiplier? (Speaking only about the CPU itself)

2011-01-25, 03:07	#539
James Heinrich "James Heinrich" May 2004 ex-Northern Ontario 7×13×47 Posts	It's now possible to directly compare between CPUs on my throughput calculator. For example, assuming identical clockspeeds, you can see how an i7-2500K neatly outruns an i7-965 by a noticable amount.