Superpi score proofs Phenom X4 is a bit faster than Athlon64 X2...

[Zucker2k]

Oh such a great reason... Like someone else said (I forget who) if running superpi all day is what you do then fine, go intel. But I know that I'm not the only one doesn't spent all day running superpi. Shoot, I don't really have much of a life and I don't sit here running pi all day. How many others then who actually have a life are going to do it all day? Seriously?

Dude, You don't have to run superpi all day to appreciate what I'm saying. Sure there are people who do that. All I'm saying is that it takes a significant amount of time (I estimate about 7 minutes, could be more) for people on AMD systems who run the 32M Pi to finish the benchmark on the average. Since Intel systems overclock better these days, this number is even significantly higher. Need I say that this also translates into "real world" apps like encoding, conversion, etc? For me performance almost always translates into time saved, not only in benchmarks but also in "real world" apps. This is the sole reason why I will switch to AMD if they become the performance king. Cheers!!

[Start]

charles i will tell you what i think feel about AMD.

I see them moving away from needing to prove they are the fastest, we have already seen tri fire on 790 with an aircooled CPU running 3d06 scoring with some god awful high score http://www.theinquirer.net/en/inquir...30000-3dmark06
which we all know has been confirmed to be true.

So an aircooled system running 3ghz beats the near 30k run I read about this week with a yorkie running on LN2, I know what i find easier to do even though im benching with 80L of LN2 and a yorkie in 14hrs to see just how far the boards will push this CPU...still for me air cooling is way easier then LN2 anyday . Now i know when crossfireX meets q9650 the scores will probably go ballistic but for now that kinda impressed me some what.

Next up, super pi...i asked you before for the source code so that i can have it ported to GPU, I really think once AMD port GPU to CPU we will see some sick fast scores that just make these CPU scores seem real slow. I remember one guy telling me a 2900XT has the power to run 32m pi in under 1minute, you know i would dearly love to see if that is true.

So you asked how i feel about AMD, I feel they are moving to a new arena, the issue is none of us really know what they are upto as none of us are privy to any real info as AMD are the kings of being tight lipped...so we just specualte and most of the time talk rubbish.

This is easily possible since a GPU can pull upwards of 500 GigaFlops, while the fastest Intel chip would be lucky to pull 90 Gigaflops (high estimate). But the problem lies in the GPU architecture, it is heavily parallized and you need to pipeline a lot of stuff in order to fully utilize the GPU. So that means having to redo some algorithms to make it properly work and exploit the full strength of the GPU. So even though you have taken the source code, redoing some of the code will take a bit of work. Then again, I haven't had experience with parallelization and pipelining yet to give a full response, but to the people familiar with the field makes this claim when I asked them

[da-key]

Interesting read here. SO let me get this straight.

Intel cpu's have pretty much always been the superpi leader. For ex. the Pentium M s479 and highly clocked p4's. AMD has never dominated there imo.

Most AMD owners in the past including myself have used spi to bench there own system speed compared to other AMD systems. I have never thought of it as a good comparison between platforms. IE Intel vs. AMD. Instead it's more like Intel vs. Intel newest release in that dept.

Overall bandwidth and latency contributed greatly to why AMD CPU's were so successful at the s939 stage when paired with ddr. The throughput for games was superior to Intel's. Also the scaling of that performance was very closely tied to CPU clock and hence with the high OC of most s939 AMD was winner at 3dmark.

My current everday gaming / rig is still an Opteron 185 s939 part that scores 11k in 3d06 with a 320mb 8800gts but only 25s SPi time. Score LINK

I feel that my system is still competitive and the score shows with a e6300 and 8800gtx losing to me even though that system probably gets 18s in SPi 1M.

So what conclusion can you draw from a SPi time? Not much other than its a prerelease bios, cpu, mobo and we don't have many other AMD Phenom X4 systems to compare it to just yet.

Enough of the Intel vs. AMD bickering. IMO AMD would have been better off to skip AM2 & ddr2 altogether. What a waste of R&D.

[n91htmare]

charles i will tell you what i think feel about AMD.

I see them moving away from needing to prove they are the fastest, we have already seen tri fire on 790 with an aircooled CPU running 3d06 scoring with some god awful high score http://www.theinquirer.net/en/inquir...30000-3dmark06
which we all know has been confirmed to be true.

So an aircooled system running 3ghz beats the near 30k run I read about this week with a yorkie running on LN2, I know what i find easier to do even though im benching with 80L of LN2 and a yorkie in 14hrs to see just how far the boards will push this CPU...still for me air cooling is way easier then LN2 anyday . Now i know when crossfireX meets q9650 the scores will probably go ballistic but for now that kinda impressed me some what.

Right, because the Tri-fire market is so Huge.

Don't you think it's sad that Intel can still break 30K mark with 2 cards? You phrased it all wrong.
It's 3ghz on air with 3 cards vs. LN2 Yorkie with 2 cards

It's nice to have the option of 3 cards, but Considering how small the SLI/Crossfire market is, the Tri-Fire will be Minuscule.

Enough of the Intel vs. AMD bickering. IMO AMD would have been better off to skip AM2 & ddr2 altogether. What a waste of R&D.

Introducing a New socket, a new memory controller on a new Architecture (K10) all at the same time could have been disastrous.

[Zucker2k]

My current everday gaming / rig is still an Opteron 185 s939 part that scores 11k in 3d06 with a 320mb 8800gts but only 25s SPi time. Score LINK

I feel that my system is still competitive and the score shows with a e6300 and 8800gtx losing to me even though that system probably gets 18s in SPi 1M.

I'm impressed! But if we have to be thruthful, lets take a brief look at the numbers:

Stock Nvidia 8800 Ultra 612/1080
Stock Nvidia 8800 GTX 575/900
Your OC'ed 8800 GTS 680/1030 (500/800 stock)

No wonder a 6300 with a GTX wouldn't touch you - you're way past Ultra territory!!!

Except maybe if they came to the party with this: http://www.xtremesystems.org/forums/...ighlight=e6300 and your oc'ed GTS

[Entity_Razer]

No......It's time to be honest and come out and say: "We have this, we sell it that much and that's it"......Don't put me snakes that will unleach their power and bite you.....Don't put me dogs/cats/etc annimals........Put me a picture of the Phenom and say: "That's it guys....if you want it, get it"......

you know hipro, your marketing skills suck

[happychappy]

Looks like all us nay-sayers were right all along

[hollo]

err, isn't that 30k 3d06 score widely considered to be total bs?
eg http://www.fudzilla.com/index.php?op...2808&Itemid=35

[n91htmare]

err, isn't that 30k 3d06 score widely considered to be total bs?
eg http://www.fudzilla.com/index.php?op...2808&Itemid=35

The notion was that it was done on 2 cards.... but it's probably done on 3 cards

[hipro5]

charles i will tell you what i think feel about AMD.

I see them moving away from needing to prove they are the fastest, we have already seen tri fire on 790 with an aircooled CPU running 3d06 scoring with some god awful high score http://www.theinquirer.net/en/inquir...30000-3dmark06
which we all know has been confirmed to be true.

Wow, wow, wow......Take it easy M8!.....
WHO and WHERE it was confirmed?.....

you know hipro, your marketing skills suck

[Shintai]

Wow, wow, wow......Take it easy M8!.....
WHO and WHERE it was confirmed?.....

My exact words aswell.

[savantu]

You can't get 30k with a single QC and a 3 GPU setup whatever magic dust is inside K10 ( less than in Kentsfield apparently ) .....

[thenrz]

You can't get 30k with a single QC and a 3 GPU setup whatever magic dust is inside K10 ( less than in Kentsfield apparently ) .....

Ahh, only a Sith talks in absolutes. Honestly, having proof works both ways. Let's see some proof that it is BS!

[informal]

You can't get 30k with a single QC and a 3 GPU setup whatever magic dust is inside K10 ( less than in Kentsfield apparently ) .....

Tony (and Fugger) obviously disagree with you on this one.And Tony said they both knew someone credible confirmed the score with the Phenom 3fire setup.

[Kasparz]

You can't get 30k with a single QC and a 3 GPU setup whatever magic dust is inside K10 ( less than in Kentsfield apparently ) .....

It might sound a little impossible about 30k, but Tony knows what he are talking about, and if he says that score is valid, then score IS valid.

[hipro5]

It might sound a little impossible about 30k, but Tony knows what he are talking about, and if he says that score is valid, then score IS valid.

At the DEFAULT resolution of the bench?...

[informal]

At the DEFAULT resolution of the bench?...

No,it ran at 640*480 .... /sarcasm

[Tony]

Wow, wow, wow......Take it easy M8!.....
WHO and WHERE it was confirmed?.....

You still have access to XIP george, i hear it was confirmed in there

Im perfectly calm m8, just pissed the Q9650 we tested today was cold bugged thats all.

[hipro5]

You still have access to XIP george, i hear it was confirmed in there

Im perfectly calm m8, just pissed the Q9650 we tested today was cold bugged thats all.

Link?....

Mobo's cold bug with York.....York uses a bit different way of ouputing data of the temp sensor below zero.... Needed a bios update for that too and NOT only a bios that will "recognise" Yorkfield....

[MaxxxRacer]

I would like to see that claim

I'll send u screenie

[Tony]

i can't link to XIP george, i just heard on the grape vine it was confirmed in there and that ther only public place i can tell you of...plus I am not welcome in XIP anymore.

Regards the yorkie, it was not my cpu, i helped James from Custom PC and Bazx and friends as they wanted to play super pi.

we hit 8.8secs but did not save it...have some 9's saved but sub 5ghz. Overall the boards were a real mess and we gave up in the end.

Still i have 30L LN2 here to play tomorrow

[KTE]

Ask various IT site owners around. Most of them said "tri-fire" the minute those results were released. I remember NH (Andreas) saying something similar too.

[hipro5]

i can't link to XIP george, i just heard on the grape vine it was confirmed in there and that ther only public place i can tell you of...plus I am not welcome in XIP anymore.

Regards the yorkie, it was not my cpu, i helped James from Custom PC and Bazx and friends as they wanted to play super pi.

we hit 8.8secs but did not save it...have some 9's saved but sub 5ghz. Overall the boards were a real mess and we gave up in the end.

Still i have 30L LN2 here to play tomorrow

About 30k on 2k6 and in VIP's section, sorry but I didn't find anything that proves/confirms it....

Ha, ha, ha......Tony......You're playing with LN2 M8?........

[mongoled]

i can't link to XIP george, i just heard on the grape vine it was confirmed in there and that ther only public place i can tell you of...plus I am not welcome in XIP anymore.

Regards the yorkie, it was not my cpu, i helped James from Custom PC and Bazx and friends as they wanted to play super pi.

we hit 8.8secs but did not save it...have some 9's saved but sub 5ghz. Overall the boards were a real mess and we gave up in the end.

Still i have 30L LN2 here to play tomorrow

-- OFFTOPIC --

Oh my, so its looking more and more that the roots upon which XS was built are slowly eroding away. Why an earth have you been excluded from the XIP Section? Oh I can definitely guess that one!

I wonder if is possible in this day and age to create a forum such as XS with has created so much traffic and for it to remain IMPARTIAL to 'outside influences' from ANY parties?

This day and age I think this is highly unlikely.

I wonder what would it take to get you re-instated in the XIP section, come on XS users, start shouting from the roof tops.

Tony In, Tony In

heheheheee

Seriously I hope whoever has thrown their DUMMY out the pram will see to reason and allow you back in......

-- END --

Back on topic, how can the title of this thread be correct?

When P4's were still blazing past K8 in superPi were they faster CPU's?

We ALL know the answer to this.

and yet still this thread is 5 pages strong and people are still discussing this

'Superpi score proofs Phenom X4 is a bit faster than Athlon64 X2...'

What gives? The thread should say

'DOES Superpi score prove that Phenom X4 is a bit faster than Athlon64 X2...????

Now we would have a thread, heheheheee

[STEvil]

Tony's welcome back if he wants I would assume, sometimes we all take ourselves and our fellow man a bit too seriously.. I think the bar's been running dry lately

As to impartiality, thats one of the top creeds of this place otherwise you wouldnt be around, you'd be with the fanboys at [H] or noobs at Tom's

Oh dear, misplaced my asbestos suit!