Akosf question: A36 vs C37?

anders n
anders n
Joined: 29 Aug 05
Posts: 123
Credit: 1,656,300
RAC: 0

RE: C37 is really slower on

Message 25842 in response to message 25838

Quote:
C37 is really slower on some intel processors (Pentium-III <5%, Pentium-M <0.3%), the cost of compatibility.
Can somebody tell me a result of AMD K6?

I have a AuthenticAMD
AMD Athlon(tm) 64 Processor 3400+ and it seem as the C37 is the same or maybe 1-2% slower than A36

I also run a P4

it runns a unit in 12500 sek now instead of 15900 with A36.

Anders n

edit [more info]

Crunchers For More Power
Crunchers For M...
Joined: 3 Aug 05
Posts: 69
Credit: 1,071,273
RAC: 0

RE: I'm working on it,

Message 25843 in response to message 25841

Quote:
I'm working on it, because of fun.

I'd like to test it if finished ;)

Thank you for your ambitious work.

CU HiNuN

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: I'd like to test it if

Message 25844 in response to message 25843

Quote:
I'd like to test it if finished ;)

I hope, you know that a SSE2 code won't be faster than optimised FPU code if your processor has just 64-bit wide SSE engine.

Scott Brown
Scott Brown
Joined: 9 Feb 05
Posts: 38
Credit: 215,235
RAC: 0

C37 running about 5-10%


C37 running about 5-10% faster than A36 on my Pentium D 830 (times are now a bit less than half those on the original albert app).

Crunchers For More Power
Crunchers For M...
Joined: 3 Aug 05
Posts: 69
Credit: 1,071,273
RAC: 0

RE: I hope, you know that a

Quote:
I hope, you know that a SSE2 code won't be faster than optimised FPU code if your processor has just 64-bit wide SSE engine.

Yes, it's clear. :)

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: RE: I hope, you know

Message 25847 in response to message 25846

Quote:
Quote:
I hope, you know that a SSE2 code won't be faster than optimised FPU code if your processor has just 64-bit wide SSE engine.

Yes, it's clear. :)

Perhaps do you have a Conroe?

Crunchers For More Power
Crunchers For M...
Joined: 3 Aug 05
Posts: 69
Credit: 1,071,273
RAC: 0

RE: Perhaps do you have a

Message 25848 in response to message 25847

Quote:
Perhaps do you have a Conroe?

No, I could test it on a Yonah-Notebook.
IMHO the Yonah processor support also 128bit-SSE2, doesn't it?

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: RE: Perhaps do you

Message 25849 in response to message 25848

Quote:
Quote:
Perhaps do you have a Conroe?
No, I could test it on a Yonah-Notebook.
IMHO the Yonah processor support also 128bit-SSE2, doesn't it?

Yonah supports SSE instructions and the SSE registers are 128-bit wide, but the execution engine is just 64-bit wide. (The processor core does 2 overlapped instructions but with some latency. It means about 20% performance loss.)

Crunchers For More Power
Crunchers For M...
Joined: 3 Aug 05
Posts: 69
Credit: 1,071,273
RAC: 0

RE: but the execution

Message 25850 in response to message 25849

Quote:
but the execution engine is just 64-bit wide.

So my thought was wrong, that Yonah already support "real" 128-bit instruction execution. Sorry. :-/

Ok, let's buy a Conroe ;)

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: Ok, let's buy a Conroe

Message 25851 in response to message 25850

Quote:
Ok, let's buy a Conroe ;)

Yeah... :)

Well, I recode the core of albert, but I cannot test it because of lack of SSE2 capabale machine. Where is my Conroe??? :)
Would you like to test it? S37a
This is a SSE2 version of C37.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.