P4, in RC5 & OGR.

[emtonsit]

OK this probably doesn't mean very much, as new cores need to be released.
Review here






<< The P4's performance with the distributed.net client ? one of the world's leading suckers up of otherwise unused CPU cycles ? was appallingly bad. My 800MHz Athlon crushed the 1.5GHz P4 by 50% for RC5 cracking (trying to break a strongly encrypted message by brute force) and by 20% for OGR (finding Optima Golomb Rulers, a task which is actually somewhat useful.) >>

 

[Train]

im gonna bet a core written for the P4 would blow away an Athlon 800, the Athlon has a very specialized RC5 core.
P4 bencmarks without its own core are meaningless

 

[TuffGuy]

well put. kind of like the p3 'crushing' the athlon/tbird/duron in seti@home. once a version that is optimized for the p4 comes out we'll truly see what it can do. btw, now that mustang has been scrapped, what does amd have planned in response to the p4?

 

[BurntKooshie]

dual Palimano's (spelling?). Mustang is still a product - just not one they plan on bringing to market (yet). If the need arises, they said they'd rethink bringing it out.

 

[TuffGuy]

palominos? sounds more like an intel chip. if they had it, i'd think that they would be eager to release it to counter the p4. my guess is that not all the kinks have been worked out yet. pity.

 

[emtonsit]

I did say that the results probably don't mean much, and I would hope that with new optimised core a 1.5GHz processor would beat a 800 MHz processor, Otherwise, if this is the future of processor design, then its pretty depressing for those doing RC5. On the subject of the Athlon having a very specialised RC5 core, there was supposed to be more improvements to come, and I doubt that it is anywhere near as specialised as the G4 RC5 core.

Is it just me or are AMD running out of decent codenames? Thoroughbred?

 

[vss1980]

Possibly AMD are running out of codenames, but Intel have run out of real-names, how many different chips can we call a Penitum-X ?

 

[emtonsit]

Good point.

More on AMD code names from AMDzone:


<< All these horse codenames are gonna get silly, I can't wait for the &quot;Mr Ed&quot; chip. >>

 

[TuffGuy]

emtonsit -

 

[jatwell]

lmao: &quot;Mr Ed&quot;

 

[ViRGE]

The P4 may truely be weak in RC5. While it's ALU is running at 2x, there's the issues of it's IPC, and if it supports rotate instruction that RC5 uses heavuly. the K5 had it, the K6 didn't, and while the K6 was faster in mhz, it was slower clock for clock than the K5; meaning it didn't pass the K5 for some time. Hopefully, it's just a core problem; but until a Dnet coder has a chance to take a look at it, we may have to live with the fact that the P4 may suck at RC5.:Q

 

[WetSprocket]

I thought the &quot;Mr Ed&quot; was the voice recognition chip.