Chipset for Socket A DOES make a difference

[nut581]

I have 2x 1700+ DLT3C JIUHB 0310MPMW (both have exact same steppings). here are how each of them being set up:

1. Abit NF7 - 11.5x200 @ 1.775Vcore - 2300Mhz - 2x256Mb DC PC3200

2. Abit KD7 - 14.0x166 @ 1.85Vcore = 2324Mhz - 1x128Mb SC PC2700

although they are exact same chip, the one on KD7 requires more voltage to get ~2300Mhz than the one on the NF7 .. what I just found out is:

they both are doing p917, the one on NF7 takes 4.48 per frame, while the one on KD7 takes 5.16 per frame.

how's that possible??????? I thought the raw mhz is all that counts? (they both have the -advmethods -forceSSE flags, and both using v4.0 official ..).

[Ducker]

Probably the higher fsb on the NF7 & the better ram. Bandwidth does have some impact on folding, not very much, but some. But I would think that it's primarly the higher fsb.

[OSUmaxx]

Are you running the same OS on both systems? Is one optimized by removing unneeded services?

[nut581]

Originally posted by OSUmaxx
Are you running the same OS on both systems? Is one optimized by removing unneeded services?

yep, both run Arkaine's overclockix, and have exact same -cfgfold settings .. :( ...

[nikhsub1]

Originally posted by nut581
I have 2x 1700+ DLT3C JIUHB 0310MPMW (both have exact same steppings). here are how each of them being set up:

1. Abit NF7 - 11.5x200 @ 1.775Vcore - 2300Mhz - 2x256Mb DC PC3200

2. Abit KD7 - 14.0x166 @ 1.85Vcore = 2324Mhz - 1x128Mb SC PC2700

although they are exact same chip, the one on KD7 requires more voltage to get ~2300Mhz than the one on the NF7 .. what I just found out is:

they both are doing p917, the one on NF7 takes 4.48 per frame, while the one on KD7 takes 5.16 per frame.

how's that possible??????? I thought the raw mhz is all that counts? (they both have the -advmethods -forceSSE flags, and both using v4.0 official ..).
Even though it is the same project number, it is most likely a different RUN of the protein. Copy the entire contents of one FAH folder to the other machine so you run the EXACT same WU, then report back with times...

[nut581]

Originally posted by nikhsub1

Even though it is the same project number, it is most likely a different RUN of the protein. Copy the entire contents of one FAH folder to the other machine so you run the EXACT same WU, then report back with times...

just did this .. p920, run 39, clone 57, gen. 1

after 5 frames, the avg on the KD7 still stays ~5.15 ... that's just 1 second faster .. and I think can be neglected .. arghhh ... it's frustrating .. 30 secs per frame = 3000 secs per WU = 50 mins :( ...

[Quailane]

Drag that out over a year and you'll really be crying.