Conroe; no threat to AMD line up...

I copied and pasted this information from the slow website. I'm not sure about the data's integrity, but still interesting reading. I read alot about the Conroe, and would have switched back to Intel based on it's performance. If you read the entire thing, you will be stunned as well. - Joeteck



Conroe performance claim being busted

Recall Intel's Mooly Eden said Con-roe will be 20% faster than AMD's future chips without even knowing AMD's plans? During the Spring 2006 IDF, Intel setup a Conroe and an Athlon 64 box, then directed benchmarkers such as Anand to push buttons*, but peaking into Windows device manager of the alleged Conroe wasn't allowed.

During the IDF, I emailed various Intel execs, AMD execs and Anand, I pointed out that such a pre-arranged blackbox Intel setup against AMD was unfair and challenged Intel to lend the Conroe box to Anand for a real drill. However, Intel dared not to answer such a simple challenge based on the rules of fair competition. The INQ sharply criticised this kind of guerilla benchmarketing.

Now, for the very first time, someone actually got hold of a Conroe chip in their own lab and did some tests. It was a 2.4GHZ Conroe CPU-Z against an Athlon 64 overclocked to 2.8GHZ. The overclocked Athlon 64 had a 2.8/2.4 -1 = 16.7% clockspeed advantage.

The following results were obtained by running 32 bit ScienceMark binaries optimized for Intel Pentium:

Molecular Dynamics
A64: 1872.68
Conroe : 2133.38 -- 14% faster

Primordia (Energy calculations for 1 atom)
Athlon64: 1506.83 -- 10% faster
Conroe: 1365.85

Cryptography
Athlon64: 1345.05 -- 26.3% faster
Conroe: 1065.59

STREAM
Athlon64: 1512.55 -- 21.7% faster
Conroe: 1242.94

The above results were for an Athlon overclocked to 2.8GHZ and a Conroe at 2.4GHZ, with the Athlon having a 16.7% clockspeed advantage. For a direct comparision at the same clockspeed, we normalize the Conroe scores by taking into account the frequency difference. Assuming the best scenario in which Conroe scores scale linearly with clock speed, we multiply the Conroe scores by a factor of 2.8/2.4. Thus, with a 2.8GHZ Conroe, we would have

Molecular Dynamics
Athlon 64 2.8GHZ: 1872.68
Conroe 2.8GHZ : 2133.38 * 2.8/2.4 = 2489 -- 32.9% faster

Primordia (Atom)
Athlon64 2.8GHZ: 1506.83
Conroe 2.8GHZ: 1365.85 * 2.8/2.4 = 1593.49 -- 5.7% faster

Cryptography
Athlon64 2.8GHZ: 1345.05 -- 8.2% faster
Conroe 2.8GHZ: 1065.59 * 2.8/2.4

STREAM *
Athlon64 2.8GHZ: 1512.55 -- 4.3% faster
Conroe 2.8GHZ: 1242.94 * 2.8/2.4 = 1450


ScienceMark is a strictly CPU/memory test, it doesn't involve video or disk I/O, it is basically a raw speed test. The ScienceMark is freely available from http://www.sciencemark.org/ for both Windows XP and Windows XP x64.

However, the above results showed a violent CPU performance fluctuation for Conroe, from it being 32% faster to being 8% slower. How can this be explained?

The cause of the Conroe performance fluctuations can't be the types of computation involved. We notice that MolDyn is a floating point computation while the Cipher is an integer computation. However, both MolDyn and Primordia are floating point calaculations on quantum mechanical properties of matter, yet, the Primodia showed a 27% relative performance drop.

As we look deeper in the ScienceMark, we notice that in the default MolDyn benchmark setting, there are only 4 cells with a simple cubic lattice, no more than 32 molecules are involved, about 2MB to 4MB memory is needed. The Primodia calculation for a single Ag (silver) atom with 47 electrons needs just a bit more memory than MolDyn. However, both the Cipher and STREAM tests involve a lot more than 4MB.

The reason why Conroe did so well in the MolDyn test is simple: Conroe has a huge 4MB of unified cache, for such single threaded tests that can fit in 4MB*, Conroe can just run off the cache with very high speed -- another cheap gimmick at the expense of very large die size.

However, once you go over the 4MB limit, Conroe is slower than Athlon 64 at the same clock. Both the Cryptography and STREM tests use a lot more than 4MB, larger than Conroe's 4MB cache, and Conroe immediately falls below Athlon 64 on the performance curve.

I can bet on this: if one increases the number of cells in the MolDyn test to 9, thus increases the working set to larger than 4MB, Conroe will perform worse than Athlon 64 at the same clockspeed.

The conclusion is: clock for clock, Athlon 64 will beat Conroe in real application environments that require a working set of larger than 4MB, or in other words, larger than Conroe's 4MB cache. This means in any real multi-tasking or server environment the Core architecture will be an underdog. Even worse, for Intel's shared cache architecture, cache thrashing is a distinct possibility under heavy loads.

I assume the Athlon 64 chip was an X2, right? Otherwise an interesting read. Still there is so much info out there you really don't know which one to believe.

I'd still love to see some SSandra scores and such.

dan

Don't think this really belongs in the AMD CPUs section.

cw823 said:

Don't think this really belongs in the AMD CPUs section.

Click to expand...

Why not.. It is what we are up against.

Joeteck said:

Why not.. It is what we are up against.

Click to expand...

Because it looks to me it may be more about Conroe, which last time I checked was an Intel processor.

GCRD

so why did you make two of the same threads in diff sections?
btw, i dont believe those results, if you look over @ XS, conroe is kicking ***

When intel previewed this chip on a compatible board, it really kicked @$$, and if you read up on Victor's progress, he's been having serious memory compatability issues and is using hardware that doesn't properly support it......

I'm not making any judgement calls until there is supported hardware, and better benchmarking stats available

This is a double post of this thread in the intel section:
http://www.ocforums.com/showthread.php?p=4427488

I'm somewhat confused trying to follow this thread around.

If not "CPU/Intel", then wouldn't "CPU/General CPU Discussion" be more appropriate?

Or did it end up here, in a "no post count" section, due to the possibility of the discussion lapsing into a fanboi flamefest?

Strat

^^^

Gotcha'!

I'm really curious to see how fast they'll get the clockspeeds to ramp up on this new arch. What I'd like to see eventually is the integration of an on-die memory controller to reduce latency for cache misses (although a 4M cache is very nice).

who wrote this up?

I believe that the guy who wrote this up doesn't have very many accurate things to say, see his website if you don't believe me.
http://sharikou.blogspot.com/2006/04/conroe-performance-claim-being-busted.html

Also he is the only person I know who thinks that AMD is going to have more Fab capacity than Intel at 65nm when AMD hasn't even produced a single 65nm chip at fab 36 and probably isn't going to produce any fab 30 for a long long time
http://sharikou.blogspot.com/2006/03/amd-poised-to-exit-2006-with-55-market.html

Also he is trying to say that AMD shouldn't deal with Dell and should instead deal with HP as it will be a win for everyone except Intel. Which makes no sense Dell dwarfs HP and currently from what am able to see(although that is somewhat limited) Dell also has a pretty decent stranglehold on severs becuase the support and build quality are better than HPs.
http://sharikou.blogspot.com/2005/12/how-can-amd-play-dell-game-for-fun-and.html

Also I can find now record of published works and no record of a doctorate thesis. So my conclusion would be AMD fanboy with to much time on his hands.

Honestly, all these benchmarks are amusing at best to me. They make for a nice read yes, though without proper hardware support and a true sample not cream of the crop samples we really won't know what these things can do. So I am just going to wait till it actually comes out and probably like 1 or 2 months after that for some mobo issues to be worked out till I start to take the results as the real deal.

I've given up following the early Conroe threads. So much speculation still. These ES Conroes are going to score differently that the final releases, and the mobos are a joke up to this point. I still think Conroe will be noticably better than AMD's offerings (tho my fanboy side hopes that doesnt happen), but until Conroe is availible to all of us I'm not going to waste my time watching it.

Ah, so THIS will be the flame war thread here. Why must this guy's drivel be posted on every forum on the Net?

speedbump said:

I believe that the guy who wrote this up doesn't have very many accurate things to say, see his website if you don't believe me.

Click to expand...

Agreed. I LOLed at a LOT of the things that guy had to say. Read other parts of is blog and you'll see just how much of an AMD fanboy he is. I mean, if you look up the word "fanatic" in the dictionary, you'll see his picture.

I never once saw any mention of the Conroe setup used to get those numbers and $10 says he has no clue how they were obtained. In fact, he likely didn't obtain them at all, rather he found them on XS or they were handed to him. As we've seen on XS, there's plenty of people with current AMDs making comparisons like that and everyone seems to skip over the fact that Conroe results there are obtained using SUPER slow DDR2 speeds (DDR-533...< half of what some of us run and there's DDR1 faster than that) with timings not even a non-OCer wouldn't use (5-5-5-15). DDR1-400 CAS3 would kill that mem on another Intel board, let alone on an AMD.

That's a very poor starting point to use for making claims like he is and don't even get me started on "bench racers" that try to normalize two different systems for comparison with simple 4th grade math.

If a fairer comparison will be made, it will be made with similar AM2/Conroe systems since they will both be DDR2 and hopefully by then, there will be beta Conroe BIOSes that finally allow ANY kind of adjustments at all. VW has no FSB, Vc or mem control at all which is why his mem is so slow (266FSB @ 1:1 = DDR2-533 and bySPD = 5-5-5-15). Conroe is on reference boards at best and not even the "enthusiast" chipset at that, so the title of this thread may be a little premature

Ross said:

Ah, so THIS will be the flame war thread here. Why must this guy's drivel be posted on every forum on the Net?
Agreed. I LOLed at a LOT of the things that guy had to say. Read other parts of is blog and you'll see just how much of an AMD fanboy he is. I mean, if you look up the word "fanatic" in the dictionary, you'll see his picture.

I never once saw any mention of the Conroe setup used to get those numbers and $10 says he has no clue how they were obtained. In fact, he likely didn't obtain them at all, rather he found them on XS or they were handed to him. As we've seen on XS, there's plenty of people with current AMDs making comparisons like that and everyone seems to skip over the fact that Conroe results there are obtained using SUPER slow DDR2 speeds (DDR-533...< half of what some of us run and there's DDR1 faster than that) with timings not even a non-OCer wouldn't use (5-5-5-15). DDR1-400 CAS3 would kill that mem on another Intel board, let alone on an AMD.

That's a very poor starting point to use for making claims like he is and don't even get me started on "bench racers" that try to normalize two different systems for comparison with simple 4th grade math.

If a fairer comparison will be made, it will be made with similar AM2/Conroe systems since they will both be DDR2 and hopefully by then, there will be beta Conroe BIOSes that finally allow ANY kind of adjustments at all. VW has no FSB, Vc or mem control at all which is why his mem is so slow (266FSB @ 1:1 = DDR2-533 and bySPD = 5-5-5-15). Conroe is on reference boards at best and not even the "enthusiast" chipset at that, so the title of this thread may be a little premature

Click to expand...

If you look closely, the AMD used is a 754 socket. Single channel memory controller. Doh!!!

If you look closely, the AMD used is a 754 socket. Single channel memory controller. Doh!!!

Click to expand...

Think my 661 isn't better at some benches than my 930? We already have a good idea how they stack up against 939 FX's. Before people start saying that the AMD system was not "ideal" or what you might run, you must also concede that neither are the Conroe benches we're seeing (or AM2 #s for that matter).

If people like to make comparisons starting with skewed data to start, that's fine. Garbage in, garbage out and we won't even mention the assumptions made to do the calculations. Regardless, they'll do what they will do, but the thread title is just flame bait and quite simply not true. Conroe won't threaten AMD in the performance arena is about the most ridiculous thing I've heard and every bit of evidence states the contrary. Whether it's 10% better on some benches, 20% better on others or -5% on some, that's a far cry from the -40% (?) on P4 Conroe *will* threaten AMD in performance like nothing has since before 2000.

Ross said:

Think my 661 isn't better at some benches than my 930? We already have a good idea how they stack up against 939 FX's. Before people start saying that the AMD system was not "ideal" or what you might run, you must also concede that neither are the Conroe benches we're seeing (or AM2 #s for that matter).

If people like to make comparisons starting with skewed data to start, that's fine. Garbage in, garbage out and we won't even mention the assumptions made to do the calculations. Regardless, they'll do what they will do, but the thread title is just flame bait and quite simply not true. Conroe won't threaten AMD in the performance arena is about the most ridiculous thing I've heard and every bit of evidence states the contrary. Whether it's 10% better on some benches, 20% better on others or -5% on some, that's a far cry from the -40% (?) on P4 Conroe *will* threaten AMD in performance like nothing has since before 2000.

Click to expand...

what I do know is that the "netburst" technology is gone. It had a 31-stage pipeline with 3 execution units. Thermal output of 115 watts.

The Conroe has the new "core" technology. Which has a 14-stage pipeline, with 4 execution units. Thermal output of 65 watts.

The conroe can process 4 intructions per clock cycle. 33% increase from netburst. Pretty awesome!

The Pentium D & P4 take two instuctions to complete a SSE instruction. The Conroe can do it in one. Plus Intel has this new technology called Macro Fusion. This can combine two instructions into one.

I just can't wait to see it.