[QUOTE=rx7350;151931]have you started LL testing exponenets yet, and, if so, what are the iteration times, and what mix of work are you running on all cores? I would be particulary interested in iteration times when running a 2560 FFT exponent (LL test) in all four cores.[/QUOTE]Sorry, it's not actually my system. I had access to run benchmarks for a few minutes, I don't actually have $4500 to spare to buy this system myself :sad:

Here's a benchmark I just ran on 64 bit Prime95 v25.9 under Vista SP1 64 bit on an i7 965 overclocked to 4.2 GHz.

However, both Vista and Prime95 get the speed wrong and say it's 4.032 GHz, which would be correct if I was running 24*168, but I'm running 25*168. CPUZ and Everest (not to mention BIOS) report the correct frequency.

Memory - 10*168 = 1680 MHz
Uncore - 20*168 = 3360 MHz
QPI - 44*168 = 7392 MHz

Edit: I should also mention that I have Turbo mode disabled so it isn't that messing up the speeds, and I've also disabled all other clock speed altering features I can find such as EIST. And since I'm already editing, I may as well mention I have HT on.

[QUOTE=lavalamp;154951]Here's a benchmark I just ran on 64 bit Prime95 v25.9 under Vista SP1 64 bit on an i7 965 overclocked to 4.2 GHz.[/QUOTE]Me want.

I would love to throw that at the 100mdpp numbers.

[quote=lavalamp;154951]However, both Vista and Prime95 get the speed wrong and say it's 4.032 GHz, which would be correct if I was running 24*168, but I'm running 25*168. CPUZ and Everest (not to mention BIOS) report the correct frequency.
[/quote]
prime95 seems to get the speed of the cpu wrong if the multiplier isnt the standard
a few days ago i underclocked my Q6600 to the minimum 6x200 and prime95 detected the cpu speed 1.8 GHz instead of 1.2
i suspect if vista got it wrong for you then prime95 gets it cpu speed from the operating system
is anyone able to try this with xp or earlier

This benchmark done with 64-bit Debian. The 64-bit Vista benchmark we performed was a little faster overall but we do not plan to use Windows. We wonder why the Linux version is slower.

We have not had time to analyze the results to determine what work this box is best for. We are just happy it easily runs Linux and that it was cheap enough to buy 2 of them.

Each box has 6GiB of PC2-6400 memory configured as 2x2GiB and 2x1GiB. We are tempted to randomly move the memory around to end up with 1 8GiB box and 1 4GiB box.

The boxes run significantly cooler compared to our older 65nm Q6600 C2Q. We think the 9550 is a 65nm CPU with a 95W TDP. We verified that the CPU has the fixed "B3" stepping (?) to eliminate the "TLB" problem.

[code]AMD Phenom(tm) 9550 Quad-Core Processor
CPU speed: 2209.26 MHz, 4 cores
CPU features: RDTSC, CMOV, Prefetch, 3DNow!, MMX, SSE, SSE2
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 48
L2 TLBS: 512
Prime95 64-bit version 25.7, RdtscTiming=1
Best time for 4K FFT length: 0.061 ms.
Best time for 5K FFT length: 0.079 ms.
Best time for 6K FFT length: 0.098 ms.
Best time for 7K FFT length: 0.126 ms.
Best time for 8K FFT length: 0.135 ms.
Best time for 10K FFT length: 0.187 ms.
Best time for 12K FFT length: 0.232 ms.
Best time for 14K FFT length: 0.276 ms.
Best time for 16K FFT length: 0.307 ms.
Best time for 20K FFT length: 0.382 ms.
Best time for 24K FFT length: 0.466 ms.
Best time for 28K FFT length: 0.564 ms.
Best time for 32K FFT length: 0.629 ms.
Best time for 40K FFT length: 0.849 ms.
Best time for 48K FFT length: 1.048 ms.
Best time for 56K FFT length: 1.249 ms.
Best time for 64K FFT length: 1.397 ms.
Best time for 80K FFT length: 1.856 ms.
Best time for 96K FFT length: 2.280 ms.
Best time for 112K FFT length: 2.726 ms.
Best time for 128K FFT length: 3.025 ms.
Best time for 160K FFT length: 3.637 ms.
Best time for 192K FFT length: 4.554 ms.
Best time for 224K FFT length: 5.691 ms.
Best time for 256K FFT length: 6.293 ms.
Best time for 320K FFT length: 7.934 ms.
Best time for 384K FFT length: 9.743 ms.
Best time for 448K FFT length: 11.669 ms.
Best time for 512K FFT length: 13.152 ms.
Best time for 640K FFT length: 17.631 ms.
Best time for 768K FFT length: 21.720 ms.
Best time for 896K FFT length: 26.135 ms.
Best time for 1024K FFT length: 29.398 ms.
Best time for 1280K FFT length: 36.024 ms.
Best time for 1536K FFT length: 44.867 ms.
Best time for 1792K FFT length: 53.138 ms.
Best time for 2048K FFT length: 60.868 ms.
Best time for 2560K FFT length: 79.030 ms.
Best time for 3072K FFT length: 96.597 ms.
Best time for 3584K FFT length: 114.752 ms.
Best time for 4096K FFT length: 129.854 ms.
Best time for 5120K FFT length: 173.594 ms.
Best time for 6144K FFT length: 214.310 ms.
Best time for 7168K FFT length: 261.567 ms.
Best time for 8192K FFT length: 298.422 ms.
Best time for 10240K FFT length: 375.800 ms.
Best time for 12288K FFT length: 458.715 ms.
Best time for 14336K FFT length: 556.367 ms.
Best time for 16384K FFT length: 645.672 ms.
Timing FFTs using 2 threads.
Best time for 4K FFT length: 0.061 ms.
Best time for 5K FFT length: 0.079 ms.
Best time for 6K FFT length: 0.098 ms.
Best time for 7K FFT length: 0.126 ms.
Best time for 8K FFT length: 0.136 ms.
Best time for 10K FFT length: 0.239 ms.
Best time for 12K FFT length: 0.226 ms.
Best time for 14K FFT length: 0.219 ms.
Best time for 16K FFT length: 0.251 ms.
Best time for 20K FFT length: 0.302 ms.
Best time for 24K FFT length: 0.356 ms.
Best time for 28K FFT length: 0.427 ms.
Best time for 32K FFT length: 0.475 ms.
Best time for 40K FFT length: 1.370 ms.
Best time for 48K FFT length: 1.270 ms.
Best time for 56K FFT length: 1.596 ms.
Best time for 64K FFT length: 1.602 ms.
Best time for 80K FFT length: 1.815 ms.
Best time for 96K FFT length: 1.701 ms.
Best time for 112K FFT length: 2.008 ms.
Best time for 128K FFT length: 2.262 ms.
Best time for 160K FFT length: 2.624 ms.
Best time for 192K FFT length: 3.176 ms.
Best time for 224K FFT length: 3.756 ms.
Best time for 256K FFT length: 4.249 ms.
Best time for 320K FFT length: 6.091 ms.
Best time for 384K FFT length: 7.367 ms.
Best time for 448K FFT length: 8.772 ms.
Best time for 512K FFT length: 9.912 ms.
Best time for 640K FFT length: 12.008 ms.
Best time for 768K FFT length: 14.679 ms.
Best time for 896K FFT length: 19.440 ms.
Best time for 1024K FFT length: 22.147 ms.
Best time for 1280K FFT length: 26.724 ms.
Best time for 1536K FFT length: 32.666 ms.
Best time for 1792K FFT length: 38.026 ms.
Best time for 2048K FFT length: 44.282 ms.
Best time for 2560K FFT length: 57.106 ms.
Best time for 3072K FFT length: 69.353 ms.
Best time for 3584K FFT length: 81.249 ms.
Best time for 4096K FFT length: 93.375 ms.
Best time for 5120K FFT length: 116.962 ms.
Best time for 6144K FFT length: 146.003 ms.
Best time for 7168K FFT length: 174.673 ms.
Best time for 8192K FFT length: 198.231 ms.
Best time for 10240K FFT length: 245.830 ms.
Best time for 12288K FFT length: 303.108 ms.
Best time for 14336K FFT length: 365.337 ms.
Best time for 16384K FFT length: 416.781 ms.
[Sat Dec 27 02:43:39 2008]
Timing FFTs using 3 threads.
Best time for 4K FFT length: 0.061 ms.
Best time for 5K FFT length: 0.080 ms.
Best time for 6K FFT length: 0.099 ms.
Best time for 7K FFT length: 0.126 ms.
Best time for 8K FFT length: 0.137 ms.
Best time for 10K FFT length: 0.209 ms.
Best time for 12K FFT length: 0.264 ms.
Best time for 14K FFT length: 0.286 ms.
Best time for 16K FFT length: 0.297 ms.
Best time for 20K FFT length: 0.351 ms.
Best time for 24K FFT length: 0.397 ms.
Best time for 28K FFT length: 0.443 ms.
Best time for 32K FFT length: 0.483 ms.
Best time for 40K FFT length: 1.376 ms.
Best time for 48K FFT length: 1.384 ms.
Best time for 56K FFT length: 1.523 ms.
Best time for 64K FFT length: 1.601 ms.
Best time for 80K FFT length: 1.828 ms.
Best time for 96K FFT length: 1.958 ms.
Best time for 112K FFT length: 1.970 ms.
Best time for 128K FFT length: 2.255 ms.
Best time for 160K FFT length: 2.416 ms.
Best time for 192K FFT length: 2.923 ms.
Best time for 224K FFT length: 3.323 ms.
Best time for 256K FFT length: 3.702 ms.
Best time for 320K FFT length: 5.783 ms.
Best time for 384K FFT length: 7.128 ms.
Best time for 448K FFT length: 8.057 ms.
Best time for 512K FFT length: 9.221 ms.
Best time for 640K FFT length: 11.140 ms.
Best time for 768K FFT length: 13.422 ms.
Best time for 896K FFT length: 21.799 ms.
Best time for 1024K FFT length: 24.119 ms.
Best time for 1280K FFT length: 27.895 ms.
Best time for 1536K FFT length: 31.639 ms.
Best time for 1792K FFT length: 35.759 ms.
Best time for 2048K FFT length: 40.462 ms.
Best time for 2560K FFT length: 58.671 ms.
Best time for 3072K FFT length: 65.984 ms.
Best time for 3584K FFT length: 74.562 ms.
Best time for 4096K FFT length: 84.000 ms.
Best time for 5120K FFT length: 91.984 ms.
Best time for 6144K FFT length: 109.437 ms.
Best time for 7168K FFT length: 129.379 ms.
Best time for 8192K FFT length: 144.422 ms.
Best time for 10240K FFT length: 191.983 ms.
Best time for 12288K FFT length: 226.256 ms.
Best time for 14336K FFT length: 267.539 ms.
Best time for 16384K FFT length: 298.965 ms.
Timing FFTs using 4 threads.
Best time for 4K FFT length: 0.061 ms.
Best time for 5K FFT length: 0.080 ms.
Best time for 6K FFT length: 0.099 ms.
Best time for 7K FFT length: 0.125 ms.
Best time for 8K FFT length: 0.136 ms.
Best time for 10K FFT length: 0.219 ms.
Best time for 12K FFT length: 0.227 ms.
Best time for 14K FFT length: 0.250 ms.
Best time for 16K FFT length: 0.271 ms.
Best time for 20K FFT length: 0.315 ms.
Best time for 24K FFT length: 0.356 ms.
Best time for 28K FFT length: 0.402 ms.
Best time for 32K FFT length: 0.446 ms.
Best time for 40K FFT length: 1.272 ms.
Best time for 48K FFT length: 1.436 ms.
Best time for 56K FFT length: 1.465 ms.
Best time for 64K FFT length: 1.269 ms.
Best time for 80K FFT length: 1.686 ms.
Best time for 96K FFT length: 1.812 ms.
Best time for 112K FFT length: 2.055 ms.
Best time for 128K FFT length: 2.160 ms.
Best time for 160K FFT length: 2.262 ms.
Best time for 192K FFT length: 2.767 ms.
Best time for 224K FFT length: 3.143 ms.
Best time for 256K FFT length: 3.452 ms.
Best time for 320K FFT length: 5.824 ms.
Best time for 384K FFT length: 6.365 ms.
Best time for 448K FFT length: 7.645 ms.
Best time for 512K FFT length: 8.738 ms.
Best time for 640K FFT length: 10.797 ms.
Best time for 768K FFT length: 12.442 ms.
Best time for 896K FFT length: 19.084 ms.
Best time for 1024K FFT length: 22.212 ms.
Best time for 1280K FFT length: 25.697 ms.
Best time for 1536K FFT length: 29.777 ms.
Best time for 1792K FFT length: 33.300 ms.
Best time for 2048K FFT length: 37.999 ms.
Best time for 2560K FFT length: 54.037 ms.
Best time for 3072K FFT length: 61.452 ms.
Best time for 3584K FFT length: 69.847 ms.
Best time for 4096K FFT length: 78.088 ms.
Best time for 5120K FFT length: 85.693 ms.
Best time for 6144K FFT length: 102.084 ms.
Best time for 7168K FFT length: 119.986 ms.
Best time for 8192K FFT length: 134.168 ms.
Best time for 10240K FFT length: 176.884 ms.
Best time for 12288K FFT length: 208.640 ms.
Best time for 14336K FFT length: 245.829 ms.
Best time for 16384K FFT length: 274.804 ms.
Best time for 58 bit trial factors: 3.621 ms.
Best time for 59 bit trial factors: 3.619 ms.
Best time for 60 bit trial factors: 3.900 ms.
Best time for 61 bit trial factors: 3.838 ms.
Best time for 62 bit trial factors: 4.551 ms.
Best time for 63 bit trial factors: 5.210 ms.
Best time for 64 bit trial factors: 6.345 ms.
Best time for 65 bit trial factors: 7.401 ms.
Best time for 66 bit trial factors: 7.348 ms.
Best time for 67 bit trial factors: 7.335 ms.
[/code]

For comparison here are the benchmarks from a dual-core Opteron 2216 at 2400MHz. Ignore the two thread FFT timings as the other core was loaded at the time and the timings were a lot worse - about 10 times worse than single thread times. I don't know why they should be worse than single core timings. It seems that for multithread benchmarks when the first core is loaded mprime tries to use that anyway and ignores the other core which is sitting idle.

[CODE]
Dual-Core AMD Opteron(tm) Processor 2216
CPU speed: 2399.43 MHz, 2 cores
CPU features: RDTSC, CMOV, Prefetch, 3DNow!, MMX, SSE, SSE2
L1 cache size: 64 KB
L2 cache size: 1 MB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 512
Prime95 64-bit version 25.8, RdtscTiming=1
Best time for 768K FFT length: 31.017 ms.
Best time for 896K FFT length: 37.082 ms.
Best time for 1024K FFT length: 41.118 ms.
Best time for 1280K FFT length: 52.565 ms.
Best time for 1536K FFT length: 64.011 ms.
Best time for 1792K FFT length: 77.507 ms.
Best time for 2048K FFT length: 86.438 ms.
Best time for 2560K FFT length: 114.380 ms.
Best time for 3072K FFT length: 139.141 ms.
Best time for 3584K FFT length: 167.903 ms.
Best time for 4096K FFT length: 187.045 ms.
Best time for 5120K FFT length: 244.509 ms.
Best time for 6144K FFT length: 301.812 ms.
Best time for 7168K FFT length: 371.209 ms.
Best time for 8192K FFT length: 441.030 ms.
Timing FFTs using 2 threads.
Best time for 768K FFT length: 316.454 ms.
Best time for 896K FFT length: 327.884 ms.
Best time for 1024K FFT length: 447.618 ms.
Best time for 1280K FFT length: 549.972 ms.
Best time for 1536K FFT length: 911.054 ms.
Best time for 1792K FFT length: 1171.418 ms.
Best time for 2048K FFT length: 1469.092 ms.
Best time for 2560K FFT length: 1376.642 ms.
Best time for 3072K FFT length: 1770.606 ms.
Best time for 3584K FFT length: 2014.975 ms.
Best time for 4096K FFT length: 4136.921 ms.
Best time for 5120K FFT length: 1728.160 ms.
Best time for 6144K FFT length: 2419.674 ms.
Best time for 7168K FFT length: 2718.081 ms.
Best time for 8192K FFT length: 2942.590 ms.
Best time for 58 bit trial factors: 3.031 ms.
Best time for 59 bit trial factors: 3.091 ms.
Best time for 60 bit trial factors: 3.086 ms.
Best time for 61 bit trial factors: 3.295 ms.
Best time for 62 bit trial factors: 3.292 ms.
Best time for 63 bit trial factors: 3.935 ms.
Best time for 64 bit trial factors: 4.609 ms.
Best time for 65 bit trial factors: 5.626 ms.
Best time for 66 bit trial factors: 6.708 ms.
Best time for 67 bit trial factors: 6.667 ms.
[/CODE]

[QUOTE=garo;155294]I don't know why they should be worse than single core timings. It seems that for multithread benchmarks when the first core is loaded mprime tries to use that anyway and ignores the other core which is sitting idle.[/QUOTE]Perhaps it is the synchronisation causing it. One core has to wait for the other to complete its assigned job before it can continue to the next part. Just a guess, I have not examined the code for P95, so I could be completely wrong [size=1](and it wouldn't be the first time)[/size].

Prime95 25.8 32-bit vs 64-bit comparison -- notice 64-bit is slightly faster. Next, I will try some system optimizations (turning off background processes) and moderate overclocking...

Core i7 965 3.20 GHz
6 GB 1600 MHz DDR3
Both run on Vista Ultimate x64

v25.8 32-bit:
[code][Sat Dec 27 11:43:22 2008]
Compare your results to other computers at [URL]http://www.mersenne.org/report_benchmarks[/URL]
Intel(R) Core(TM) i7 CPU 965 @ 3.20GHz
CPU speed: 3200.19 MHz, 4 hyperthreaded cores
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2, SSE4
L1 cache size: 32 KB
L2 cache size: 256 KB, L3 cache size: 8064 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 64
Prime95 32-bit version 25.8, RdtscTiming=1
Best time for 768K FFT length: 12.788 ms.
Best time for 896K FFT length: 15.274 ms.
Best time for 1024K FFT length: 17.836 ms.
Best time for 1280K FFT length: 22.780 ms.
Best time for 1536K FFT length: 27.810 ms.
Best time for 1792K FFT length: 33.225 ms.
Best time for 2048K FFT length: 36.987 ms.
Best time for 2560K FFT length: 46.829 ms.
Best time for 3072K FFT length: 57.208 ms.
Best time for 3584K FFT length: 68.374 ms.
Best time for 4096K FFT length: 77.713 ms.
Best time for 5120K FFT length: 98.463 ms.
Best time for 6144K FFT length: 118.721 ms.
Best time for 7168K FFT length: 145.209 ms.
Best time for 8192K FFT length: 160.652 ms.
Timing FFTs using 2 threads on 1 physical CPUs.
Best time for 768K FFT length: 11.872 ms.
Best time for 896K FFT length: 14.180 ms.
Best time for 1024K FFT length: 16.208 ms.
Best time for 1280K FFT length: 20.046 ms.
Best time for 1536K FFT length: 24.350 ms.
Best time for 1792K FFT length: 29.039 ms.
Best time for 2048K FFT length: 32.203 ms.
Best time for 2560K FFT length: 42.506 ms.
Best time for 3072K FFT length: 51.819 ms.
Best time for 3584K FFT length: 62.680 ms.
Best time for 4096K FFT length: 69.727 ms.
Best time for 5120K FFT length: 89.028 ms.
Best time for 6144K FFT length: 107.188 ms.
Best time for 7168K FFT length: 128.920 ms.
Best time for 8192K FFT length: 142.922 ms.
Timing FFTs using 4 threads on 2 physical CPUs.
Best time for 768K FFT length: 7.895 ms.
Best time for 896K FFT length: 8.904 ms.
Best time for 1024K FFT length: 12.328 ms.
Best time for 1280K FFT length: 10.773 ms.
Best time for 1536K FFT length: 12.900 ms.
Best time for 1792K FFT length: 15.009 ms.
Best time for 2048K FFT length: 16.992 ms.
Best time for 2560K FFT length: 21.732 ms.
Best time for 3072K FFT length: 26.537 ms.
Best time for 3584K FFT length: 31.891 ms.
Best time for 4096K FFT length: 35.505 ms.
Best time for 5120K FFT length: 45.393 ms.
Best time for 6144K FFT length: 54.511 ms.
Best time for 7168K FFT length: 65.562 ms.
Best time for 8192K FFT length: 72.602 ms.
Timing FFTs using 6 threads on 3 physical CPUs.
Best time for 768K FFT length: 6.726 ms.
Best time for 896K FFT length: 7.468 ms.
Best time for 1024K FFT length: 10.840 ms.
Best time for 1280K FFT length: 8.474 ms.
Best time for 1536K FFT length: 9.869 ms.
Best time for 1792K FFT length: 11.500 ms.
Best time for 2048K FFT length: 12.927 ms.
Best time for 2560K FFT length: 15.482 ms.
Best time for 3072K FFT length: 18.988 ms.
Best time for 3584K FFT length: 21.906 ms.
Best time for 4096K FFT length: 24.980 ms.
Best time for 5120K FFT length: 31.035 ms.
Best time for 6144K FFT length: 37.737 ms.
Best time for 7168K FFT length: 45.081 ms.
Best time for 8192K FFT length: 50.929 ms.
Timing FFTs using 8 threads on 4 physical CPUs.
Best time for 768K FFT length: 6.061 ms.
Best time for 896K FFT length: 6.728 ms.
Best time for 1024K FFT length: 9.826 ms.
Best time for 1280K FFT length: 7.452 ms.
Best time for 1536K FFT length: 11.827 ms.
Best time for 1792K FFT length: 10.110 ms.
Best time for 2048K FFT length: 14.350 ms.
Best time for 2560K FFT length: 13.819 ms.
Best time for 3072K FFT length: 16.615 ms.
Best time for 3584K FFT length: 19.140 ms.
Best time for 4096K FFT length: 21.856 ms.
Best time for 5120K FFT length: 26.682 ms.
Best time for 6144K FFT length: 32.361 ms.
Best time for 7168K FFT length: 38.490 ms.
Best time for 8192K FFT length: 44.472 ms.
Best time for 58 bit trial factors: 2.987 ms.
Best time for 59 bit trial factors: 3.064 ms.
Best time for 60 bit trial factors: 3.069 ms.
Best time for 61 bit trial factors: 3.053 ms.
Best time for 62 bit trial factors: 3.071 ms.
Best time for 63 bit trial factors: 5.129 ms.
Best time for 64 bit trial factors: 5.131 ms.
Best time for 65 bit trial factors: 4.704 ms.
Best time for 66 bit trial factors: 4.687 ms.
Best time for 67 bit trial factors: 4.652 ms.
[/code]v25.8 64-bit:
[code][Sat Dec 27 11:33:44 2008]
Compare your results to other computers at [URL]http://www.mersenne.org/report_benchmarks[/URL]
Intel(R) Core(TM) i7 CPU 965 @ 3.20GHz
CPU speed: 3200.25 MHz, 4 hyperthreaded cores
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2, SSE4
L1 cache size: 32 KB
L2 cache size: 256 KB, L3 cache size: 8064 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 64
Prime95 64-bit version 25.8, RdtscTiming=1
Best time for 768K FFT length: 12.456 ms.
Best time for 896K FFT length: 15.104 ms.
Best time for 1024K FFT length: 17.513 ms.
Best time for 1280K FFT length: 22.572 ms.
Best time for 1536K FFT length: 27.685 ms.
Best time for 1792K FFT length: 33.148 ms.
Best time for 2048K FFT length: 36.760 ms.
Best time for 2560K FFT length: 46.268 ms.
Best time for 3072K FFT length: 57.111 ms.
Best time for 3584K FFT length: 68.143 ms.
Best time for 4096K FFT length: 77.640 ms.
Best time for 5120K FFT length: 98.536 ms.
Best time for 6144K FFT length: 118.232 ms.
Best time for 7168K FFT length: 144.138 ms.
Best time for 8192K FFT length: 159.407 ms.
Timing FFTs using 2 threads on 1 physical CPUs.
Best time for 768K FFT length: 11.425 ms.
Best time for 896K FFT length: 13.818 ms.
Best time for 1024K FFT length: 15.805 ms.
Best time for 1280K FFT length: 19.891 ms.
Best time for 1536K FFT length: 24.300 ms.
Best time for 1792K FFT length: 29.181 ms.
Best time for 2048K FFT length: 31.378 ms.
Best time for 2560K FFT length: 42.715 ms.
Best time for 3072K FFT length: 52.190 ms.
Best time for 3584K FFT length: 62.650 ms.
Best time for 4096K FFT length: 70.092 ms.
Best time for 5120K FFT length: 89.259 ms.
Best time for 6144K FFT length: 106.710 ms.
Best time for 7168K FFT length: 129.246 ms.
Best time for 8192K FFT length: 142.946 ms.
Timing FFTs using 4 threads on 2 physical CPUs.
Best time for 768K FFT length: 7.728 ms.
Best time for 896K FFT length: 8.404 ms.
Best time for 1024K FFT length: 11.493 ms.
Best time for 1280K FFT length: 10.701 ms.
Best time for 1536K FFT length: 12.917 ms.
Best time for 1792K FFT length: 15.146 ms.
Best time for 2048K FFT length: 16.926 ms.
Best time for 2560K FFT length: 21.580 ms.
Best time for 3072K FFT length: 26.561 ms.
Best time for 3584K FFT length: 32.208 ms.
Best time for 4096K FFT length: 35.634 ms.
Best time for 5120K FFT length: 45.296 ms.
Best time for 6144K FFT length: 55.814 ms.
Best time for 7168K FFT length: 65.725 ms.
Best time for 8192K FFT length: 72.016 ms.
Timing FFTs using 6 threads on 3 physical CPUs.
Best time for 768K FFT length: 6.531 ms.
Best time for 896K FFT length: 7.380 ms.
Best time for 1024K FFT length: 10.415 ms.
Best time for 1280K FFT length: 8.337 ms.
Best time for 1536K FFT length: 10.010 ms.
Best time for 1792K FFT length: 11.288 ms.
Best time for 2048K FFT length: 12.873 ms.
Best time for 2560K FFT length: 15.652 ms.
Best time for 3072K FFT length: 18.779 ms.
Best time for 3584K FFT length: 21.855 ms.
Best time for 4096K FFT length: 24.839 ms.
Best time for 5120K FFT length: 30.963 ms.
Best time for 6144K FFT length: 37.708 ms.
Best time for 7168K FFT length: 45.105 ms.
Best time for 8192K FFT length: 50.962 ms.
Timing FFTs using 8 threads on 4 physical CPUs.
Best time for 768K FFT length: 5.860 ms.
Best time for 896K FFT length: 6.621 ms.
Best time for 1024K FFT length: 9.411 ms.
Best time for 1280K FFT length: 7.297 ms.
Best time for 1536K FFT length: 8.735 ms.
Best time for 1792K FFT length: 9.884 ms.
Best time for 2048K FFT length: 14.211 ms.
Best time for 2560K FFT length: 13.596 ms.
Best time for 3072K FFT length: 16.441 ms.
Best time for 3584K FFT length: 18.911 ms.
Best time for 4096K FFT length: 21.624 ms.
Best time for 5120K FFT length: 26.377 ms.
Best time for 6144K FFT length: 32.239 ms.
Best time for 7168K FFT length: 38.523 ms.
Best time for 8192K FFT length: 44.362 ms.
Best time for 58 bit trial factors: 2.365 ms.
Best time for 59 bit trial factors: 2.451 ms.
Best time for 60 bit trial factors: 2.449 ms.
Best time for 61 bit trial factors: 2.702 ms.
Best time for 62 bit trial factors: 2.848 ms.
Best time for 63 bit trial factors: 3.381 ms.
Best time for 64 bit trial factors: 4.199 ms.
Best time for 65 bit trial factors: 4.441 ms.
Best time for 66 bit trial factors: 4.655 ms.
Best time for 67 bit trial factors: 4.623 ms.
[/code]

[QUOTE=Freightyard;155333]Prime95 25.8 32-bit vs 64-bit comparison -- notice 64-bit is slightly faster.[/QUOTE]

That may be due to Nehalem allowing macro-ops fusion for 64-bits instructions, whereas Conroe would only do it on 32-bits. 64-bit code may in general benefit slightly from the use of Nehalem.

Here are the runs for my Fujitsu Lifebook N series

In Safe Mode:[code]Intel(R) Core(TM)2 Duo CPU T8300 @ 2.40GHz
CPU speed: 2393.99 MHz, 2 cores
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: 3072 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 256
Prime95 32-bit version 25.6, RdtscTiming=1
Best time for 4K FFT length: 0.055 ms.
Best time for 5K FFT length: 0.081 ms.
Best time for 6K FFT length: 0.096 ms.
Best time for 7K FFT length: 0.116 ms.
Best time for 8K FFT length: 0.120 ms.
Best time for 10K FFT length: 0.169 ms.
Best time for 12K FFT length: 0.208 ms.
Best time for 14K FFT length: 0.249 ms.
Best time for 16K FFT length: 0.263 ms.
Best time for 20K FFT length: 0.353 ms.
Best time for 24K FFT length: 0.428 ms.
Best time for 28K FFT length: 0.530 ms.
Best time for 32K FFT length: 0.553 ms.
Best time for 40K FFT length: 0.722 ms.
Best time for 48K FFT length: 0.875 ms.
Best time for 56K FFT length: 1.069 ms.
Best time for 64K FFT length: 1.141 ms.
Best time for 80K FFT length: 1.609 ms.
Best time for 96K FFT length: 1.959 ms.
Best time for 112K FFT length: 2.362 ms.
Best time for 128K FFT length: 2.512 ms.
Best time for 160K FFT length: 3.086 ms.
Best time for 192K FFT length: 3.795 ms.
Best time for 224K FFT length: 4.525 ms.
Best time for 256K FFT length: 5.012 ms.
Best time for 320K FFT length: 6.514 ms.
Best time for 384K FFT length: 8.070 ms.
Best time for 448K FFT length: 9.708 ms.
Best time for 512K FFT length: 10.843 ms.
Best time for 640K FFT length: 14.443 ms.
Best time for 768K FFT length: 17.739 ms.
Best time for 896K FFT length: 21.136 ms.
Best time for 1024K FFT length: 24.438 ms.
Best time for 1280K FFT length: 29.876 ms.
Best time for 1536K FFT length: 36.451 ms.
Best time for 1792K FFT length: 43.330 ms.
Best time for 2048K FFT length: 48.106 ms.
Best time for 2560K FFT length: 63.450 ms.
Best time for 3072K FFT length: 77.505 ms.
Best time for 3584K FFT length: 92.166 ms.
Best time for 4096K FFT length: 103.026 ms.
Best time for 5120K FFT length: 131.929 ms.
Best time for 6144K FFT length: 159.328 ms.
Best time for 7168K FFT length: 193.605 ms.
Best time for 8192K FFT length: 212.122 ms.
Best time for 10240K FFT length: 284.213 ms.
Best time for 12288K FFT length: 347.446 ms.
Best time for 14336K FFT length: 421.633 ms.
Best time for 16384K FFT length: 463.288 ms.
Best time for 20480K FFT length: 633.442 ms.
Best time for 24576K FFT length: 773.084 ms.
Best time for 28672K FFT length: 924.279 ms.
Best time for 32768K FFT length: 1016.381 ms.
Timing FFTs using 2 threads.
Best time for 4K FFT length: 0.055 ms.
Best time for 5K FFT length: 0.081 ms.
Best time for 6K FFT length: 0.096 ms.
Best time for 7K FFT length: 0.116 ms.
Best time for 8K FFT length: 0.120 ms.
Best time for 10K FFT length: 0.157 ms.
Best time for 12K FFT length: 0.147 ms.
Best time for 14K FFT length: 0.162 ms.
Best time for 16K FFT length: 0.190 ms.
Best time for 20K FFT length: 0.218 ms.
Best time for 24K FFT length: 0.252 ms.
Best time for 28K FFT length: 0.306 ms.
Best time for 32K FFT length: 0.321 ms.
Best time for 40K FFT length: 0.926 ms.
Best time for 48K FFT length: 0.961 ms.
Best time for 56K FFT length: 1.071 ms.
Best time for 64K FFT length: 1.106 ms.
Best time for 80K FFT length: 1.263 ms.
Best time for 96K FFT length: 1.274 ms.
Best time for 112K FFT length: 1.332 ms.
Best time for 128K FFT length: 1.540 ms.
Best time for 160K FFT length: 1.686 ms.
Best time for 192K FFT length: 1.997 ms.
Best time for 224K FFT length: 2.373 ms.
Best time for 256K FFT length: 2.636 ms.
Best time for 320K FFT length: 3.444 ms.
Best time for 384K FFT length: 4.265 ms.
Best time for 448K FFT length: 5.134 ms.
Best time for 512K FFT length: 5.769 ms.
Best time for 640K FFT length: 7.888 ms.
Best time for 768K FFT length: 9.724 ms.
Best time for 896K FFT length: 11.591 ms.
Best time for 1024K FFT length: 13.628 ms.
Best time for 1280K FFT length: 16.233 ms.
Best time for 1536K FFT length: 19.710 ms.
Best time for 1792K FFT length: 23.539 ms.
Best time for 2048K FFT length: 26.324 ms.
Best time for 2560K FFT length: 34.174 ms.
Best time for 3072K FFT length: 41.476 ms.
Best time for 3584K FFT length: 49.373 ms.
Best time for 4096K FFT length: 55.409 ms.
Best time for 5120K FFT length: 70.696 ms.
Best time for 6144K FFT length: 86.309 ms.
Best time for 7168K FFT length: 104.079 ms.
Best time for 8192K FFT length: 116.549 ms.
Best time for 10240K FFT length: 148.466 ms.
Best time for 12288K FFT length: 181.124 ms.
Best time for 14336K FFT length: 218.482 ms.
Best time for 16384K FFT length: 244.767 ms.
Best time for 20480K FFT length: 341.957 ms.
Best time for 24576K FFT length: 424.301 ms.
Best time for 28672K FFT length: 494.244 ms.
Best time for 32768K FFT length: 542.375 ms.
Best time for 58 bit trial factors: 4.309 ms.
Best time for 59 bit trial factors: 4.310 ms.
Best time for 60 bit trial factors: 4.307 ms.
Best time for 61 bit trial factors: 4.302 ms.
Best time for 62 bit trial factors: 7.304 ms.
Best time for 63 bit trial factors: 7.326 ms.
Best time for 64 bit trial factors: 6.777 ms.
Best time for 65 bit trial factors: 6.749 ms.
Best time for 66 bit trial factors: 6.734 ms.
Best time for 67 bit trial factors: 6.716 ms.[/code]

Standard mode:[code]Intel(R) Core(TM)2 Duo CPU T8300 @ 2.40GHz
CPU speed: 2393.74 MHz, 2 cores
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: 3072 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 256
Prime95 32-bit version 25.6, RdtscTiming=1
Best time for 4K FFT length: 0.055 ms.
Best time for 5K FFT length: 0.081 ms.
Best time for 6K FFT length: 0.095 ms.
Best time for 7K FFT length: 0.107 ms.
Best time for 8K FFT length: 0.110 ms.
Best time for 10K FFT length: 0.155 ms.
Best time for 12K FFT length: 0.190 ms.
Best time for 14K FFT length: 0.230 ms.
Best time for 16K FFT length: 0.243 ms.
Best time for 20K FFT length: 0.324 ms.
Best time for 24K FFT length: 0.396 ms.
Best time for 28K FFT length: 0.489 ms.
Best time for 32K FFT length: 0.511 ms.
Best time for 40K FFT length: 0.667 ms.
Best time for 48K FFT length: 0.807 ms.
Best time for 56K FFT length: 0.987 ms.
Best time for 64K FFT length: 1.054 ms.
Best time for 80K FFT length: 1.485 ms.
Best time for 96K FFT length: 1.807 ms.
Best time for 112K FFT length: 2.183 ms.
Best time for 128K FFT length: 2.318 ms.
Best time for 160K FFT length: 2.849 ms.
Best time for 192K FFT length: 3.504 ms.
Best time for 224K FFT length: 4.176 ms.
Best time for 256K FFT length: 4.643 ms.
Best time for 320K FFT length: 6.025 ms.
Best time for 384K FFT length: 7.492 ms.
Best time for 448K FFT length: 9.023 ms.
Best time for 512K FFT length: 10.107 ms.
Best time for 640K FFT length: 13.472 ms.
Best time for 768K FFT length: 16.695 ms.
Best time for 896K FFT length: 19.929 ms.
Best time for 1024K FFT length: 23.071 ms.
Best time for 1280K FFT length: 28.338 ms.
Best time for 1536K FFT length: 34.502 ms.
Best time for 1792K FFT length: 44.119 ms.
Best time for 2048K FFT length: 49.067 ms.
Best time for 2560K FFT length: 64.672 ms.
Best time for 3072K FFT length: 78.852 ms.
Best time for 3584K FFT length: 87.954 ms.
Best time for 4096K FFT length: 98.101 ms.
Best time for 5120K FFT length: 125.420 ms.
Best time for 6144K FFT length: 154.790 ms.
Best time for 7168K FFT length: 188.334 ms.
Best time for 8192K FFT length: 202.024 ms.
Best time for 10240K FFT length: 269.291 ms.
Best time for 12288K FFT length: 325.543 ms.
Best time for 14336K FFT length: 394.450 ms.
Best time for 16384K FFT length: 446.469 ms.
Best time for 20480K FFT length: 600.071 ms.
Best time for 24576K FFT length: 735.307 ms.
Best time for 28672K FFT length: 884.776 ms.
Best time for 32768K FFT length: 969.536 ms.
Timing FFTs using 2 threads.
Best time for 4K FFT length: 0.050 ms.
Best time for 5K FFT length: 0.074 ms.
Best time for 6K FFT length: 0.088 ms.
Best time for 7K FFT length: 0.107 ms.
Best time for 8K FFT length: 0.110 ms.
Best time for 10K FFT length: 0.167 ms.
Best time for 12K FFT length: 0.158 ms.
Best time for 14K FFT length: 0.160 ms.
Best time for 16K FFT length: 0.182 ms.
Best time for 20K FFT length: 0.213 ms.
Best time for 24K FFT length: 0.255 ms.
Best time for 28K FFT length: 0.307 ms.
Best time for 32K FFT length: 0.322 ms.
Best time for 40K FFT length: 1.070 ms.
Best time for 48K FFT length: 1.103 ms.
Best time for 56K FFT length: 1.212 ms.
Best time for 64K FFT length: 1.249 ms.
Best time for 80K FFT length: 1.435 ms.
Best time for 96K FFT length: 1.350 ms.
Best time for 112K FFT length: 1.337 ms.
Best time for 128K FFT length: 1.488 ms.
Best time for 160K FFT length: 1.691 ms.
Best time for 192K FFT length: 2.005 ms.
Best time for 224K FFT length: 2.382 ms.
Best time for 256K FFT length: 2.654 ms.
Best time for 320K FFT length: 3.457 ms.
Best time for 384K FFT length: 4.324 ms.
Best time for 448K FFT length: 5.229 ms.
Best time for 512K FFT length: 5.934 ms.
Best time for 640K FFT length: 8.139 ms.
Best time for 768K FFT length: 10.042 ms.
Best time for 896K FFT length: 11.957 ms.
Best time for 1024K FFT length: 13.995 ms.
Best time for 1280K FFT length: 16.918 ms.
Best time for 1536K FFT length: 20.502 ms.
Best time for 1792K FFT length: 24.036 ms.
Best time for 2048K FFT length: 27.105 ms.
Best time for 2560K FFT length: 34.617 ms.
Best time for 3072K FFT length: 42.533 ms.
Best time for 3584K FFT length: 50.500 ms.
Best time for 4096K FFT length: 57.145 ms.
Best time for 5120K FFT length: 73.119 ms.
Best time for 6144K FFT length: 89.457 ms.
Best time for 7168K FFT length: 108.792 ms.
Best time for 8192K FFT length: 120.360 ms.
Best time for 10240K FFT length: 153.076 ms.
Best time for 12288K FFT length: 186.586 ms.
Best time for 14336K FFT length: 226.672 ms.
Best time for 16384K FFT length: 259.571 ms.
Best time for 20480K FFT length: 347.952 ms.
Best time for 24576K FFT length: 423.961 ms.
Best time for 28672K FFT length: 515.468 ms.
Best time for 32768K FFT length: 551.724 ms.
Best time for 58 bit trial factors: 4.299 ms.
Best time for 59 bit trial factors: 3.983 ms.
Best time for 60 bit trial factors: 3.972 ms.
Best time for 61 bit trial factors: 3.975 ms.
Best time for 62 bit trial factors: 6.792 ms.
Best time for 63 bit trial factors: 7.342 ms.
Best time for 64 bit trial factors: 6.257 ms.
Best time for 65 bit trial factors: 6.256 ms.
Best time for 66 bit trial factors: 6.219 ms.
Best time for 67 bit trial factors: 6.223 ms.[/code]

Running in safemode I get an actual performance gain much larger.
~2200 seconds between outputs in safemode vs. ~3300 seconds in standard.

how do you get it to do that huge benchmark