HT and all kind of other optimisation thread

[hhh]

This thread is ment to gather and provide information on how to make the best use of every kind of computer system.

I just enabled Hyperthreading on my P4 3.0GHz, on Linux and did some tests. Here my numbers. (192K FFT, 100T)
LLR3.6 alone _6.413 ms/bit
LLR/sieve____12.831 ms/bit___ 179kp/s
sieve alone_________________179 kp/s
2 x llr_____12.907 ms/bit (x2)
2 x sieve____________________180 kp/s (x2)

That means that the sieve client takes the best advantage of HT, giving the P4 a not that bad sieve performance, IMO. In fact, when you run a sieve client, running something else is free.
The LLR client (Primality testing) does not take any advantage of two instances.

Does this fit with your experiences? What about AMD? And many other possible questions.
H.

[Citrix]

http://www.mersenneforum.org/showthr...ighlight=hyper

[hhh]

Originally Posted by Citrix

http://www.mersenneforum.org/showthr...ighlight=hyper

This URL seem to me just to consider prime95, but neither the LLR-client nor sieving (at least these word don't show up). H.