Is it me or does amd xp cpu's pummel p4's in SorB?

[Esquire]

my 1700+ is faster than a p4 2.53?

 

[DannyBoy]

It will be.

The P4 takes advantage of SSE2 instructions in the SOB client, and seeing as the recent bug has disabled them in the latest client they are crap for SOB until the client is fixed.

 

[Overkiller]

The SSE2 bug in SOB is very crippling. Before a 2.8 ghz P4 @ 3.0 would do 700+ K cems/s

Now my 2.8 is only a 1/3 faster than my 1.2 ghz athlon :Q

The bug should be fixed soon as the developers are awaiting the end of their finals (hehe) to have time to work this out.

 

[Allio]

Yup. It's a complete reversal of what it used to be. Don't write it off for P4s yet, the bug will be fixed at some stage

 

[InverseOfNeo]

Its just taking them a LONG time.

 

[Overkiller]

The man has finals...please be understanding

But yeah..you'd think for such a large project he'd well...yeah :/

 

[ken008]

MMX instuctions were barely used because they were hard for the programer to implement. SSE2 may suffer from the same problem.

 

[Confused]

Originally posted by: ken008
MMX instuctions were barely used because they were hard for the programer to implement. SSE2 may suffer from the same problem.

SSE2 was already implemented, and it gave GREAT advantages to the P4. However, once n got to 498xxx or something, there was a bug that caused any SSE2 enabled CPUs to crash out. This is why 1.1.1 was released, disabling SSE2 for P4s and Opterons (but not Pentium-M CPUs ). Without SSE2, the P4 is crippled, hence why a much slower Athlon XP currently races waaay ahead.

Once Louis' finals are over, then he should hopefully be able to get the SSE2 bug fixed, and release another version, which will put P4s back at the top of the peformance charts


Garry

 

[BlackMountainCow]

Hehe, yet another reason for AMD only