Coolaler @ 11.266s w/ 4.53Ghz Conroe !!!

[***Deimos***]

With the 2.93, I'm sure 5ghz will be possible with LN2, just a matter of waiting now .

Aren't they using LN2 already o_0?
Oh and my humble congratulation to

Coolair


and

TAM.


Ps: you certainly should have made the warning about wearing a diaper before visiting this thread more visible!

[coolaler]

my host is down
so reload pic

[***Deimos***]

coolaler, could you please explain why the cycle time (TRAS) is sometimes 1 and sometimes 4.

[ak_47_boy]

OMFG, OMFG, OMFG that is f***ing fast

[coolaler]

coolaler, could you please explain why the cycle time (TRAS) is sometimes 1 and sometimes 4.

because sometimes I set 1,sometimes unset

[MAS]

coolaler
try http://athlon64venice.narod.ru/MAS.rar on core 2

[coolaler]

coolaler
try http://athlon64venice.narod.ru/MAS.rar on core 2

thank u

[k0nsl]

Aren't they using LN2 already o_0?

Um, yes, but on ES processors and with motherboards that isn't exactly the best.

-k0nsl

[SoF]

I wonder what the little program compares...not even at 5GHZ an AMD would break the superpi times of your beast

[MAS]

in superpi both data/stack and code segments are in 4M L2 cache
reading and writing data therefore are extremly fast

look at 16k ... 256k superpi conroe results - it's not so good in comparison to K8

[M.Beier]

in superpi both data/stack and code segments are in 4M L2 cache
reading and writing data therefore are extremly fast

look at 16k ... 256k superpi conroe results - it's not so good in comparison to K8

Compare 32M ...
4ghz A64 wont even get close..

[stone_cold_Jimi]

Compare 32M ...
4ghz A64 wont even get close..

I still want to see 2 instances of SPi run. I'm betting it won't be as fast, but not far off, judging by my reading of the cache architecture. But no-one seems interested in running them

Coolaler, be the first to publish dual SPi results m8. Then all us AMD owners can thoughtfully stroke our chins and say "hmmm, now I understand..." lol.

[MAS]

I still want to see 2 instances of SPi run.

two instances of superpi are 10-15% slower than one

[MAS]

I wonder what the little program compares

comment: based on electronic signature on elliptic curves algorythms

[Clairvoyant129]

thank u

A 3.8GHz Venice? Ok... sure

[Big SturL]

This thread has made me let Intel into the heat again. May I ask, what was the original factoryset clock for this CPU?

[Gamer]

2.66 Ghz.

[BlackX]

So why CPU-Z shows 9x multi ??
266MHz x 10 = 2,67 GHz

[MAS]

imho mult. was specially reduced to find the best operating mode (both cpu and memory have max clockspeed)

[M.Beier]

Speedstep works I guess

[SoF]

comment: based on electronic signature on elliptic curves algorythms

well that wasn't what I asked for but nevertheless - I never liked comparing AMD to Intel that way at all Real benches are the true way to compare performance for me.

[[cTx]Philosophy]

hehe, Whats he running? A P4 @ 1.7ghz? Time for an upgrade...

even worse p4 @ 1.5 256M crapolla...

[***Deimos***]

Im not quite sure how SuperPi works, but the similarity in K8 to Conroe scores at <<1M is probably because of overhead, and granularity of measuring. ie both scores are very small numbers.

I am anxiously waiting for SuperPi32M results. Although, I'm not expecting any surprised. Others posted SuperPI 32M with Conroe and results look spectacular - even at stock clock speeds. Quite frankly, with something as powerful, and efficient as Conroe, I'm not sure why you'd need to overclock, and perhaps Intel is digging their own grave here. Afterall, with something this phenomenal, how could they ever persuade Conroe users to upgrade.

I think its also a pretty tought time to sell used P4 stuff.. who would want it? To make matters worse, Intel is slashing prices like crazy to empty P4 inventory.

EDIT: Do you think, if you could travel back in time, it would be more impressive or cool to show P2/Celeron/P3/K62 400 people of the year 1998: an AthlonXP@400Mhz, a AM2 K8 (dual core)@400Mhz, Conroe @400Mhz, P4@400Mhz, or Itanium2@400Mhz. Showing P4 (especially P4 D) would certainly be interesting because of radically longer pipeline, and trace cache and other novel architecture differences. Showing K8 and telling folks back then that someday Intel will cower and tremble before the might of AMD could get you a fast-pass to the nut house. Conroe might turn out to be actually least exciting because of its first glance similarities to P6 (in comparison to P4). Revealing that after billions of R&D Intel would only sell a handfull of their much hyped Itanium might be a shocker to some... just some.

[***Deimos***]

even worse p4 @ 1.5 256M crapolla...

Hey, remember the P4 1.3Ghz (Williamete 256KB L2, with PC133 SDRAM). Blazing fast P4 for OEM right? In order to cut down the size of the P4, Intel had to make a lot of concessions (256KB L2 and SDRAM says it all). It was a hard sell getting those early P4's out to knowledgable consumers when even the 1.5Ghz version didn't always beat a 1Ghz P3. The big problem back then was Intel using i850 RAMBUS performance to hype up performance, but customers unable to afford it getting lower end chipsets... afterall its just the chipset - how much difference could it possibly make, right? Throughout subsequent years it seemed it didn't matter, because sockets, chipsets, and VRM requirements quickly changed.

Kudos to Intel with Conroe:
- Intel finally returns to the IPC race, it seemed to abandon in 2001.
- uses existing, at least a bit mature, socket and chipset.
- uses existing readily available DDR2
- and hopefully, judging by their agressive P4 inventory sell off, there will be sufficient Conroe available to meet demand.

PS: Can't wait to see what marketing guys have in store for us. For once they actually have the better product out there, and dont need to lie or exagerate. Weird!