CPU Hype-O-Meter

Mephistopheles

Distinguished
Feb 10, 2003
2,444
0
19,780
Just to see how hyped up you guys are about the new CPUs from Intel and AMD... It's actually an interesting poll!

Here it is: <A HREF="http://forumz.tomshardware.com/community/modules.php?name=Forums&file=viewtopic&p=33030#33030" target="_new">CPU Hype-O-Meter</A>

<font color=red><b>M</b></font color=red>ephistopheles
 
I'd think people would post moderate opinions... but so far, two votes went to "A64 will destroy Prescott" and one to "Prescott will obliterate A64"... At least one person will end up disappointed...

<font color=red><b>M</b></font color=red>ephistopheles
 
Keeping everyone busy while they wait for A64 an Prescott.
Thanks! That way, we wouldn't notice time passing by... :smile: Anyway, I'm on my holiday time now, and I have to find something interesting to do with <i>my</i> time... And while no other option presents itself (like the beach or something... I'm working on that!... nothing like living in a tropical climate) I will just keep writing nonsense on this forum... sorry, everyone! :wink:

<font color=red><b>M</b></font color=red>ephistopheles
 
Well, I have a few hobbies. I like swimming as a sport, I do some weightlifting, I play piano and I mess with computers. Also, the government granted me a scholarship to study physics, so I have to study at least a bit of physics, or they'll be giving me money for no reason! And that would make me feel guilty... :frown:

Anyway, it's not as if I live by the beach. And it's not as if I forgot how to have fun, either... 😎

<font color=red><b>M</b></font color=red>ephistopheles
 
It's one of my hobbies. Plus, I'm actually into simulations right now. So while I wait for my simulations to finally come to an end, I usually use a web browser. :smile:

And granted, we're currently in our winter, and it's cold as hell - at least for our standarts.

<font color=red><b>M</b></font color=red>ephistopheles
 
Pretty cool poll Meph...
I must say though these folks voting for A64 to totally blow away Scotty are really not thinking straight. If Opteron is not blowing away Woody right now, why would we expect a64 to be a revolutionary step above. It cant happen! It will be somewhere in the middle; on par with Scotty, most likely below. The hype is definately building though. :smile:

Wanted: Large breasted live-in housekeeper. Must be a good cook, organized, and willing to pick up after me.
 
If Opteron is not blowing away Woody right now, why would we expect a64 to be a revolutionary step above. It cant happen! It will be somewhere in the middle; on par with Scotty, most likely below.
I was thinking that way too. The most reasonable options were the ones in the middle; I wouldn't expect any dramatic outcome! So in that sense, I agree with this...
I must say though these folks voting for A64 to totally blow away Scotty are really not thinking straight.
Yes, agreed. :smile:

Anyway, I think the poll is interesting because we can feel how much hype is in the air...

<font color=red><b>M</b></font color=red>ephistopheles
 
ROFL @ ChipDeath

I completely agree. I don't even see how it will overtake Scotty when running optimized 64-bit software, not to mention regular old 32-bit software which will likely run rather ... like an equally-clocked AXP.

"<i>Yeah, if you treat them like equals, it'll only encourage them to think they <b>ARE</b> your equals.</i>" - Thief from <A HREF="http://www.nuklearpower.com/daily.php?date=030603" target="_new">8-Bit Theater</A>
 
32-bit software which will likely run rather ... like an equally-clocked AXP.
Shocking. 😱 But it's probably true that it will just tickle Intel's CPUs at 32-bit, at best, not kick them in the rear end.

BTW, that <i>was</i> a good one, ChipDeath! :lol:

<font color=red><b>M</b></font color=red>ephistopheles
 
It WILL run better in 32-bit per-clock, there's no denying that, it has the memory controller and we have seen what it can do per clock. Some will go unaffected or be slightly better but overall IT HAS to be better per clock.

--
<A HREF="http://www.lochel.com/THGC/html/news.html" target="_new"><font color=purple><b>The official Tom's Hardware Guide Forums Photo Album, click here to contribute!</font color=purple></b></A>
 
It WILL run better in 32-bit per-clock, there's no denying that, it has the memory controller and we have seen what it can do per clock. Some will go unaffected or be slightly better but overall IT HAS to be better per clock.
It 'HAS' to be because you want it to, or it 'HAS' to be because of hardware, or what?

Theoretically, if we can believe AMD that the ondie memory controller helps out soooo much, it <i>might</i> perform faster than an AXP that's decked-out.

Realisticly, just how much would the slice off of the memory latency of a <i>single channel</i> memory controller help? The A64 has a slightly longer pipeline. AMD's CPUs are not exactly known for stunning prediction and prefetching (which compounds the added pipeline length because you'll have to fetch the hard way more often). The nForce2 can already run XMS, LL, HyperX, etc. at rather low latencies already and further trims memory latencies in dual-channel over anyone else's single-channel so far.

Will an AXP with an nForce2 and LL RAM <i>really</i> be slower than an A64 in pure 32-bit? Or will it run at exactly the same speed if not a teeny-tiny bit faster?

Unless AMD pulls a really nifty rabbit out of their hat (like better prediction algorithms or an ability to process more ops per cycle) for the release of the A64, I just don't see it beating an AXP at 32-bit. Short of something up a sleave, SSE2 so far is A64's only possible hope to impress. But depending on how AMD implemented SSE2 it might not actually boost performance much if any.

We'll have to wait and see, but I'm not expecting much out of AMD.

"<i>Yeah, if you treat them like equals, it'll only encourage them to think they <b>ARE</b> your equals.</i>" - Thief from <A HREF="http://www.nuklearpower.com/daily.php?date=030603" target="_new">8-Bit Theater</A>
 
Slvr, the Opteron differs so little at the moment over the A64, except using Dual Channel. Assume that gives it 10% performance. Look around the Workstation charts, or on sites that tested a very old and bad A64 sample. Everything points to a 20%+ boost, and it shows.

You did debate this so many times in the past about how it is possible to content or reach ODMC performance with the current normal Northbridge, but the fact is, they try and they try, and they simply can't. The ODMC has proven to stretch the bar significantly, and there are enough benches in the world to prove it.
I dunno and no offense, but I am somehow surprised you are being this blind! It's been on the web forever, the benches showing how single Opterons per-clock are much better. Suppose the MAXIMUM difference between A64 and Opteron in single CPU mode is 15%, just to be generous. A64 is still ahead of the AXP by all means.

PS: Also take note Opteron uses ECC. That makes an even more interesting showing for Athlon 64's relative performance.
--
<A HREF="http://www.lochel.com/THGC/html/news.html" target="_new"><font color=purple><b>The official Tom's Hardware Guide Forums Photo Album, click here to contribute!</font color=purple></b></A><P ID="edit"><FONT SIZE=-1><EM>Edited by Eden on 07/15/03 06:45 PM.</EM></FONT></P>
 
Unless AMD pulls a really nifty rabbit out of their hat (like better prediction algorithms or an ability to process more ops per cycle) for the release of the A64, I just don't see it beating an AXP at 32-bit.
Huh? We already know that per clock, an Opteron handily beats an AXP at 32-bit code. Unfortunately AMDZone pulled the benchies of the NForce3 and will update with a new revision of the board, but as I remember in most CPU-bound benchmarks the 1.8 GHz Opie was around 20-30% faster than a 1.8 GHz AXP, and much more than that in the memory-bound benchies. However, it was behind the P4 3.2C in nearly all tests.

Now it's true that the A64 looks to only have single channel memory, although recent rumors claim that the A64 1 MB version will be dual-channel capable. But it'll probably support DDR400 versus the Opie's DDR333. So overall, the A64's performance should be in the same ballpark as the Opie's. Which is significantly faster per clock than the AXP at 32-bit code.

Ritesh
 
That's exactly what I was saying, and here Anandtech <A HREF="http://www.anandtech.com/cpu/showdoc.html?i=1818&p=3" target="_new">http://www.anandtech.com/cpu/showdoc.html?i=1818&p=3</A> tries to predict A64 performance in the two pages and throws in a cache prediction comparison, knowing A64 may use 512KB L2 (or was it 256KB?). The difference is too nil.
Dual Channel is the main difference though one must wonder just HOW USEFUL is dual channel to Opteron really? Is Hyper-Transport a truly different method of transfering data and WILL make the K8 bandwidth-hungry? Who knows. Personally I just see Opteron as being 10% ahead in most benchmarks in overall. But A64 IS DEFINITELY going to hurt the AXP in 32-bit, there is just no question about it.

--
<A HREF="http://www.lochel.com/THGC/html/news.html" target="_new"><font color=purple><b>The official Tom's Hardware Guide Forums Photo Album, click here to contribute!</font color=purple></b></A>
 
However, it was behind the P4 3.2C in nearly all tests.
I suppose Intel will really make sure A64 has a hard time... That is why AMD will probably get the initial A64 out @ 2.0Ghz, and they might be considering a dual channel memory architecture after all... I guess everyone is doing their best then... That's always good for the end user. :smile:

<font color=red><b>M</b></font color=red>ephistopheles
 
Interesting link there. However, when linking to that review, it's always good to remember that PCI-ed video cards are much better on Opteron than on the P4 because of the different peripheral architectures. (Opteron uses PCI X). Moving the video card to an AGP slot will yield a big bonus to the P4 platform, and a smaller one for the Opteron platform.
Is Hyper-Transport a truly different method of transfering data and WILL make the K8 bandwidth-hungry?
That is an interesting question. We've seen and heard a lot of hype about hypertransport, but what can it deliver and what are its implications?... We'll see that later...

<font color=red><b>M</b></font color=red>ephistopheles
 
However, when linking to that review, it's always good to remember that PCI-ed video cards are much better on Opteron than on the P4 because of the different peripheral architectures. (Opteron uses PCI X). Moving the video card to an AGP slot will yield a big bonus to the P4 platform, and a smaller one for the Opteron platform.
Issue is I linked to it simply for the "Prediction Athlon 64's performance" section, not for any comparison of performance.

As for PCI or whatnot, what exactly is PCI X to PCI in performance? If both used 133MB/sec then where does it make a difference, for the sake of debating a bit of graphics involvment in CPU comparison?

That is an interesting question. We've seen and heard a lot of hype about hypertransport, but what can it deliver and what are its implications?... We'll see that later...
So far no one has dared say Opteron's "FSB". I'd like to know just how do the circuitry inside the core work with something no longer refered as FSB. Methinks HT is still an FSB by simply using serial technology for excellent scaling. But that would imply Opteron uses 800MHZ effective data rate like a P4, and if K7 is an indication, then that is useless really, as even a 400MT AthlonXP barely benefits from extra FSB and memory. So I am compelled to wonder just how HT works in this FSB thing.

--
<A HREF="http://www.lochel.com/THGC/html/news.html" target="_new"><font color=purple><b>The official Tom's Hardware Guide Forums Photo Album, click here to contribute!</font color=purple></b></A>