First post, by Carlos S. M.
I once stumbled with the idea of clocking each P4 model (per generation) at the same clock (like 1.5/1.6/1.7 ghz) and then bench it. This will be a long term project and it might take some time to materialize (due to work/studies, getting necesary parts, life...)
The plataform testbed will be based in two Asrock motherboards, the P4i65G and the 775i65G, both based on the Intel 865G and these mobos has practically the same specs except for the CPU support and Socket
The conterders will be:
Socket 478:
A Willamette P4 sample (got some samples, planning to get an 1.6 Ghz Willamette just in case)
A Northwood P4 sample (got some samples, planning to get an 3.4 Ghz and 1.7 Ghz Northwood just in case)
A Prescott P4 sample (got some samples)
A Galaltin P4 EE sample (if is possible to get one)
LGA 775:
A Prescott-1M P4 sample (got some samples)
A Prescott-2M P4 sample (got some samples)
A Cedar Mill P4 sample (got a P4 631, might try getting a 641/651 as well)
CPU Clock Options (depending of the models i got):
1.5 Ghz (100x15)
1.6 Ghz (100x16)
1.8 Ghz (100x17)
Northwood/Prescott/Cedar Mill parts will be benched with both HT on and off, 64 bit P4s will be benched in 32 and 64 bit as well.
I'm not benching any of the Celeron varaints due to multiplier options in them (only 1-2 Celerons has a x17 multi), also not benching the LGA 775 version of the Gallatin due to begin extremenly rare
Other hardware
Asrock P4i65G for Socket 478
Asrock 775i65G R1/R2 for LGA 775 (still have to get one)
2 GB DDR 400
AGP videocard (probably using my 7600GT or my HD 4650)
SATA HDD
Software and benchmarks:
Windows XP SP3 (32 bit tests) and XP x64 SP2 (64 bit tests)
Passmark PerfomanceTest
7-Zip
Clibench
3DMark series (like 99, 2000, 2001, 03, 05, 06...)
other benchmark suggestions?
¿What do you think about the idea of benching each P4 gen clock at the same clock? ¿What P4 gen/core would likely win overall?
I'll try to keep this update as long i get the reaming parts and start working on this