V24.12 special benchmarks needed - Page 3

Prime95 · 2005-06-14, 17:55

Quote:

Originally Posted by Cruelty

But P3 supports "prefetch"... so how come the error became visible

When you choose "AllBench=1" it benchmarks all implementations of the FFT, including those it would not normally choose to run on your computer.

The oddball timings pointed to a typo in the entry for the 1024k all-complex no-prefetch FFT.

delta_t · 2005-06-14, 19:04

here's another pentium m
this one is from a the new 2ghz dothan with 533mhz bus, and ddr2 ram. dell inspiron 6000 laptop.

Cruelty · 2005-06-14, 20:20

Quote:

Originally Posted by Prime95

When you choose "AllBench=1" it benchmarks all implementations of the FFT, including those it would not normally choose to run on your computer.

Thanks for the info. I guess it's good that bug has been discovered anyway

hhh · 2005-06-16, 02:18

Voila:
Intel(R) Celeron(R) CPU 2.00GHz
CPU speed: 1991.82 MHz
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 8 KB
L2 cache size: 128 KB
L1 cache line size: 64 bytes
L2 cache line size: 128 bytes

Yours H.

lunna · 2005-06-30, 10:33

Pentium D (P4 Prescott dual-core)

This was run according to the instructions. Second core was idle during the test.

lunna · 2005-06-30, 10:56

Pentium D (P4 Prescott dual-core)

Pentium D 830 (3.0GHz clocked at 3.5GHz)
ASUS P5WD2 Premium/Wifi-TV/D
2GB DDR2 PC4200 ECC memory (2*1GB)

When the above run (previous post) was finished, I re-run the test but with CPU affinity set.

My results and outlnder's (3rd post) should be comparable as we have the same speed, but my Pentium D is faster. There are some variants within a FFT size that are performing better than the ordinary Prescott. If that is because of the i955X chipset, ASUS tuning, or Intel tuning the cores is anyones guess.

outlnder · 2005-06-30, 18:14

I'm betting the difference comes from the throughput between processor and memory. The standard Prescott only has 478 pins while the 830 D has 775. Thus allowing a faster throughput to memory and back.

The real test would be to compare a 478 Prescott and 775 Prescott running at the same speed with similar motherboards and Ram.

lunna · 2005-07-04, 09:59

Don't know why, but I just assumed it was the 775 version. Anyway, I seem to recall that the 775 pin socket wasn't about enhancing bandwidth, only giving more power to the CPU (and moving the pins to the socket of course).

The difference between our machines is probably due to the chipset or memory type (or speed).

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
AMD benchmarks NO LONGER needed for v26	Prime95	Software	11	2012-01-13 15:06
Celeron - special benchmarks needed	Prime95	Software	9	2011-04-07 02:19
Special cluster analysis needed?	Mystwalker	Miscellaneous Math	2	2005-09-10 00:13
Celeron D Benchmarks needed	E_tron	Hardware	4	2004-08-10 11:28
v23.5 benchmarks needed	Prime95	Software	44	2003-08-07 06:57

2005-06-30, 18:14	#29
outlnder Aug 2002 2×3×53 Posts	I'm betting the difference comes from the throughput between processor and memory. The standard Prescott only has 478 pins while the 830 D has 775. Thus allowing a faster throughput to memory and back. The real test would be to compare a 478 Prescott and 775 Prescott running at the same speed with similar motherboards and Ram.

2005-07-04, 09:59	#30
lunna Mar 2003 2⁴ Posts	Don't know why, but I just assumed it was the 775 version. Anyway, I seem to recall that the 775 pin socket wasn't about enhancing bandwidth, only giving more power to the CPU (and moving the pins to the socket of course). The difference between our machines is probably due to the chipset or memory type (or speed).