AMD Phenom vs INTEL C2D and C2Q - 3D rendering and gaming benchmarks

totally useless

since you can over clock an amd quad past 3ghz

you should only only overclock an intel qaud to 3.6ghz

stock speeds are uselss since even dell and hp oc now


WarpedSystems only sells overclocked systems, since 2003!

I have to question that benchmark something seems off the test i did on phenom's were much lower and results don't match?.

I don't know if it's been all the lament from enthusiasts but that link really gives an idea that stock vs. stock performance isn't all that bad.

I wonder what will happen to this architecture if they have a decent 45nm process?

exactly. But it's not exactly tainted once removing it's OC capabilities.

I'll have to tell my GF and grandparents the big news. They no longer have to submit themselves to the horrible rendering times that the Phenoms promote. Their days will be so much more productive. score!

While spreading the good news in my Paul Revere clothing I'll be sure to pee on this author's car door handle. That way it's inescapable. If he wants to go home he has to touch it!!

Really? Well, perhaps you would care to share some portion of those volumes?

This was never in doubt in my mind -- AMD does a good job pricing their processors slightly lower than intel's performance -- at stock/binned speeds. And that's a fair comparison.

The problem with this is that we still have not yet heard rumours of 45nm Phenom having sucessfully booted. Remember, it takes about 1 year to go from first silicon to saleable product. The clock is ticking.

Also, note that AMD's 45nm is not in the same league as intel's 45nm -- since they are still using an oxide gate dielectric. Take a look at the minimal transistor performance between 90nm and 65nm and you'd expect to see about the same for the 45nm transition. End story -- all other things being equal, AMD will still be competitive clock for clock, but not with thermals, and their clock headroom is seriously impacted.

Have you seen this benchmark at LostCircuits (http://www.lostcircuits.com/cpu/amd_phenom/20.shtml)?
Strangely the Phenoms win out on UT3, and I tend to agree with the opinion that Phenom has some sort of unaccounted "power" which isn't being taken advantage of on current applications.
Interesting.

The good thing about phenom is how it kicks my Athlon 4400 and whenever I get the money I will replace the Athlon and go phenom.
I could go wolfie with the new asrock+wifi which I find hot, but nobody will buy my actual AM2 asrock.
I don't oc, but I tried to push my Athlon to 2.9Ghz stable without raising the voltage. (but I have to admit the asrock has a slightly higher than standar core voltaje)

Meh, we all know Im just an evil paid Intel fanboy pumper thats just here to insitgate trouble and dilute facts

I have seen the Phenom 9500 priced around the £130 range on some sites and when combined with a cheap motherboard that is good value.

I have issues with this report. 1, it included ES AMD procs. 2, nowhere in the preface (or indeed anywhere else that I could find) did it mention pricing as an objective. 3, following up on 1 and 2, it did NOT include the qx9650 or the qx6850.

Also, the tests chosen were (outside of Sandra and autodesk tests) relatively limited in terms of being comparable, and even sandra and autodesk are not all that widely used. Why not SPEC?

So, from a buyer's perspective, this report is useless. I can't buy ES chips, and AMD can't claim that production chips would be as good. (Tahoe, anyone?) Tests are not very well accepted, nor widely used. No standardised tests were conducted. Production quads were left out of the sample mix.

I think I smell a 'slant'.

If by 'holds its own' you mean 'loses out in almost every real application benchmark to a lower-clocked Intel quad core', yes.

We know that Phenom wins on many synthetic benchmarks because of the integrated memory controller; we also know that that makes little difference in the real world for desktop PCs, except in a few specialised applications.

I don't mind seeing the benchmarks, but then, the only part of the article that bugs me:

Heh... should be faster. I'm really starting to hate hyped up articles. It would be nice to see B3's out there now for these types of articles.

But then, at least they did mention a mixed of retail phenoms 9500, 9600.

It was an alright read, in my opinion... but it doesn't change the fact that I have an Intel system. . o O (too bad AMD, i would have had a AMD system if it was affordable back then)

^Agreed.

Pardon me for bytching, but I'm running a s939 DFI with an opty 175... I found that this article left much to be desired when it comes to making a real decision as to upgrading. I have fairly deep pockets, can afford almost anything I want. However my Scottish ancestry tells me that I need to be thrifty, but not to the point of absurdity.

Now that you've got me in a rant mode, AMD left all of the s939 users out to dry. The newer AM2 boards are more expensive, and their Intel counterparts are becoming very competitive. Memory prices being the same, tell me again why should I go with AMD? To get a lesser quality processor that uses more watts? To save 100AUD up front, only to pay higher electric bills over the next four years? BTW, I can build an Intel based system here in AUS for ~50 difference if I pick the right parts. So this is one AMD fanboy that is crying.

STFU.

how come i havnt seen phenom benched on a am2 chipset? its always benched on the 790fx chipsets... kind of throws a wrench in the whole cheaper is better campaign

"ProE results are interesting as the Athlon64 X2 5200+ was actually the fastest processor tested"

http://www.techspot.com/review/84- [...] page9.html

which version of ProE ?

is this an indication that ProE is single or dual threaded & not
using the 4 cores. or just a coincidence.

No? Didn't like the memory comment thing?

Okay... how about we benchmark all of the new 45ns Intel chips in a $30.00 Foxconn motherboard. After all it is "compatible".... and it uses DDR memory so you can save even MORE money!

Ahem, you also have to think about the fact that on the current run of Phenoms, they're crippled by the Northbridge/IMC being locked at 1.8ghz. And since the Hyper transport speed can't be any higher than the NB speed, that locks that lower than it should be as well.

As soon as Im absolutely sure I'm stable at 2.7Ghz I'm gonna up the NB to 2.6 and HTT to 2.6 to see what kind of effect that will have.

And? What relevance does it have to the fact that in these benchmarks that the AMD fanboys are touting, a Phenom at 2.6GHz is beaten in the majority of cases by a Core 2 at 2.4GHz?

Is that really supposed to be better 'clock for clock'?

Have you even bothered to read the article that this thread is about? The Q6600 beats the 2.6GHz Phenom in almost every real application benchmark in the article.

Intel releases their new chips and Phenom is still an also-ran.

Excuse me? Try to publish something on 45nm within the past year.

I work in the semiconductor industry, so if you wanted to contradict me, you would have pointed to the just-released IEDM transistor data. Not some write up of promises from IBM's team. Folks, it ain't pretty for IBMs process compared to Intel's.

http://www.realworldtech.com/page. [...] 222300&p=6

Look for table 1. IBM needs to implement hiK/MG ASAP.

well, considering the fact that IBM has already tested and done a press release on their hiK/MG it says a lot. But they've tested theirs at 32nm. AMD has nudged them into adapting it to 45nm for k10 45nm update.

Try again. Show me a working testchip or core. Ask anybody in the industry about what IBM's "me too" press release was ~1year ago and you will get a rolling of eyes.

IBM is also not the only company working on 32nm. It is true that there are rumours of a 45mn rev2 process from IBM where the hiK is scheduled to be implemented. Truthfully, the ROI on that is limited at best. You're better off doing (as I suspect IBM will do) focusing your new R&D efforts on getting it up and running on a new node (ie, 32nm) than doing it for production on an existing node. Otherwise you'll be spreading your research engineers very thin and by the time you have it perfected and ramped your next node will already be here. Now, IBM would be stupid not to have pilot lines of Hik/MG on their 45nm node (or even 65nm node!) to get the hang of doing it such that the learning on the 32nm node won't be quite so bad. But ramping for volume production? I'm not so sure.

Nevertheless, it is certainly not an "advantage" to delay implementation to the 32nm node, nor a "technological achievement" over Intel, since they have already demonstrated a 32nm test chip (with hik/mg) back at the IDF in October. It's instead a tacit admission of being a generation behind.

In the end, press releases mean nothing. And that includes intels IDF showings. Functional chips and device data mean everything. And right now, if you would bother to look at the data, the IBM consortium (as well as the rest of the industry) is at a significant technological disadvantage to intel's 45nm process. There is no debate over this.

Acutally, the transistors for 65nm are *slower* than those of 90nm. The physics tells you that.

Looking at the IEDM data, IBM 45nm should show *some* improvement over IBM 65nm. It's hard to say how much. But it's clear that it won't be in the same league at the same thermals as intels.