SPCR

At least according to benchmarks done by intel.

Here's a link from Anandtech.

Everything looked valid according to Anandtech. Cannot wait for impartial tests... but it looks like Intel will pick the performance crown again in 6 months.

I kinda disappointed about the media benchmarks though. Conroe seems to beat the Athlon in games, but I'm more interested in media encoding performance formy HTPC. I don't plan on upgrading my HTPC until next year so I'll just have to wait and see what the playing field will be like.

stupid wrote:I kinda disappointed about the media benchmarks though. Conroe seems to beat the Athlon in games, but I'm more interested in media encoding performance formy HTPC. I don't plan on upgrading my HTPC until next year so I'll just have to wait and see what the playing field will be like.

What dissapointing about the media encoding performance? Conroe beats AMD in it too with clear numbers. For example in Divx 6.1 31 seconds vs 44 seconds with fx-60 overclocked to 2.8ghz.

Of course these are the benchmarks that Intel has handpicked, but even considering that, it looks very promising. AMD better have some ace in their sleeve. Judging from the marketing numbers by Intel I gathered that, A64s will still be cooler then Conroes, but not by a big margin.

This has been discussed on another thread. And how do you know its a FX-60 ? They only provide you with a POST screenshot showing "Unknown AMD Processor". And in the graphs they typed "Athlon 64 X2 @ 2.8GHz". So who's is it? FX-60, X2 @ 2.8GHz or Unknown ?

Intel did provide the testbed and benchmarks, I'll wait for something less "controversial".

Erssa wrote:What dissapointing about the media encoding performance? Conroe beats AMD in it too with clear numbers. For example in Divx 6.1 31 seconds vs 44 seconds with fx-60 overclocked to 2.8ghz.

Oops. I guess I was scanning it too quickly, I was thinking in terms of FPS, not time to finish.

rpsgc wrote:This has been discussed on another thread. And how do you know its a FX-60 ? They only provide you with a POST screenshot showing "Unknown AMD Processor". And in the graphs they typed "Athlon 64 X2 @ 2.8GHz". So who's is it? FX-60, X2 @ 2.8GHz or Unknown ?

Intel did provide the testbed and benchmarks, I'll wait for something less "controversial".

The AMD benchmarks aren't even that essential. It's the benchmarks of the Conroe that really matter here. It would be pretty useless for Intel to cheat on these, as every major hardware site is going to give them a fair benchmarking before they hit the market.

There is cheating, and there is cheating. nVIDIA and ATI both "cheated" a bit in benchmarks and games in the past, a little "optimization" here and there

Yeah, I wouldn't put it past Intel to fudge the numbers a bit -- not like this would be the first time.

But if those results are genuine, and representative of Conroe performance in real-life situations, then AMD isn't going to stand a chance at the end of the year. An overclocked FX-60 (essentially an FX-62), AMD's future flagship CPU getting slaughtered by a high-end Conroe? Just imagine how the Conroe XE, which is rumored to be 1333MHz FSB and upwards of 3GHz (I've heard as high as 3.33GHz), would perform. AM2, 65nm, none of these will be able to save AMD. Short of an architectural change, I can't imagine what they'd do to close the gap.

I'd be extremely skeptical of any benchmark put out by any company. Intel in particular has been notorius in the past for having put in "optimizations" in code that makes their processers run faster and AMD's run slower.
That being said, if it is true, it looks like that will be one sweet chip!

My question is what the heck has AMD been doing in the background? They can't just be sitting on their laurels and it's fair to say that the move to AM2 & Socket F (e.g. DDR2 memory) won't provide a huge performance increase. Looking at the fact that Intel's vNext processor is beating what is essentially an AMD vPrevious (technology-wise) processor, I can't believe AMD doesn't have something up its sleeve (maybe the 65nm processors?). As long as AMD has the memory controller on-die and Intel doesn't, it'll always have a huge architectual advantage over Intel.

Unless, of course, they really are that oblivious.

-Derek

rpsgc wrote:This has been discussed on another thread. And how do you know its a FX-60 ? They only provide you with a POST screenshot showing "Unknown AMD Processor". And in the graphs they typed "Athlon 64 X2 @ 2.8GHz". So who's is it? FX-60, X2 @ 2.8GHz or Unknown ?

Intel did provide the testbed and benchmarks, I'll wait for something less "controversial".

The scores are essentially the same for FEAR as compared to a 2.8GHz FX-57 with the same video card setup.

http://www.anandtech.com/video/showdoc.aspx?i=2679&p=12

I only know of the K8L but that's a server-side CPU AFAIK. Quad-core, etc. Not much of innovation there

AMD will continue to do well the same way they always have, selling chips at prices well under equivelant intel chips. Also AMD is further along on the "Performance per Watt" road than intel (unless intel wises up and starts using their mobil chips for their desktop lineup, or at least its architecture).

its all a bunch of intel propaganda at this point IMO anyhow. everyone knows you dont trust benchmarks for a product from that products manufacturer.

AMD will continue to do well the same way they always have, selling chips at prices well under equivelant intel chips.

Intel's cheapest dual-core is about £100 cheaper than AMD's cheapest dual-core here in the UK.

unless intel wises up and starts using their mobil chips for their desktop lineup, or at least its architecture).

Which is exactly what they have done with Conroe:

The Intel Core Microarchitecture is Intel's new processor architecture. Announced in Q1 2006, to be released later in the year, it will replace the old NetBurst and Pentium M microarchitectures.

The architecture features low power usage, multiple cores, virtualization, 64-bit technologies, and hyperthreading in the higher end models.

The first processors that will use this architecture are code-named Merom, Conroe, and Woodcrest; Merom is for mobile computing, Conroe is for desktop systems, and Woodcrest, which is expected to feature the most cores and technologies of all those processors, is for servers.

wiki

jaganath wrote:

AMD will continue to do well the same way they always have, selling chips at prices well under equivelant intel chips.

Intel's cheapest dual-core is about £100 cheaper than AMD's cheapest dual-core here in the UK.

Just becuase it is cheaper doesnt mean that it is equivalent. AFAIK it is far inferior to the X2 3800.

-- writing from IDF...

The comparison benchmarks were actually performed by Anand himself at IDF, in public, more or less, with lots of witnesses. It seems to have been fairly done. I wasn't present, but chatted with those who were.

A couple of obvious points --

-- Conroe will not be available for a few months
-- By then, new AMD processors on their new socket will also be available
-- The comparison was interesting and exciting, but it's naturally biased. The Intel chip was an advanced yet unreleased model, while the AMD was a current, retail model. When Intel current best is compared w/ AMD's current best, then it will be a more relevant contest.

accord1999 wrote:

rpsgc wrote:This has been discussed on another thread. And how do you know its a FX-60 ? They only provide you with a POST screenshot showing "Unknown AMD Processor". And in the graphs they typed "Athlon 64 X2 @ 2.8GHz". So who's is it? FX-60, X2 @ 2.8GHz or Unknown ?

Intel did provide the testbed and benchmarks, I'll wait for something less "controversial".

The scores are essentially the same for FEAR as compared to a 2.8GHz FX-57 with the same video card setup.

http://www.anandtech.com/video/showdoc.aspx?i=2679&p=12

Good find, that means that at least the AMD benchmarks are real.

Now a question for all you sceptics. It's pretty obvious how a graphics card can cheat on a benchmark to gain more frames, but please tell how a processor would do the same thing, when it cannot directly alternate the image quality resulting in greater fps?

As the AMD benchmark results are what you should be expecting from it, it really limits the ways Intel could cheat. Just take those media encoding benchmarks, there really isn't a way Intel can cheat there. The only way they can "cheat" is by hand picking the benchmarks, which they most certainly have done here. But despite it, it can be seen that numbers are clearly for Intel, even if they were from benchmarks that would give the most positive image about Intel, the results are nothing short of amazing.

As for the gaming benchmarks... Amd got the numbers that were expected from it.

Now the only way for Intel to cheat on this is to have somehow altered the coding of the games to actually run on a lower quality. Those guys at Anandtech, Hexus etc. that have published their articles aren't idiots. They would have probably noticed, if the cheating was something as obvious as running the tests at different quality settings. Or Intels bench having overclocked GFX. But this kind of cheating would be rediculous as it would obviosly be revealed as soon as the HW sites got their test chips. Intel would only make a fool out of themselves by doing something as blatant. And if Intel would have really done some optimization to every game code, all I can say is shit, release that code to public, if it makes the game run so much smoother...

These benchies have to be taken with a grain of salt but still... If you think intel is not going to be the performance leader after Conroes are released, you have to be a fanboys. There is no way that moving to ddr2 will grant AMD 40% performance boost in F.E.A.R. And even 40% wouldn't be enough for AMD to keep nr.1 I'd be suprised even if DDR2 gave a 5% boost.

But let's not forget that these chips aren't gonna be available for another 6 months. AMD should cherish these last moments as the manufacturer of the better processor.

Greetings,

Was it an FX60 (i.e. dual core), or was it an FX57; which IIANM is single core?

A small but important point: we do not know what the power consumption of Conroe is, and anyway, that will probably change from what it is now, to when the production chips are sold. And, we do not know what improvements AMD will make, with the AM2 socket and DDR2 RAM, with either clock speed improvements and/or architectural changes...

Will Intel have an answer for Hypertransport? How will their implimentation of 64bits compare? (Were either/both of these benchmarks 64bit, or were they differently compiled?) I know that AMD has already doubled the registers from 8 to 16 -- what did Intel do in Conroe? What about memory latency? Has Intel put the memory controller on the CPU?

Has Intel put the memory controller on the CPU?

No; I'm not even sure if that's in their roadmap.

Overview of Conroe

NeilBlanchard wrote:Greetings,

Was it an FX60 (i.e. dual core), or was it an FX57; which IIANM is single core?

A small but important point: we do not know what the power consumption of Conroe is, and anyway, that will probably change from what it is now, to when the production chips are sold. And, we do not know what improvements AMD will make, with the AM2 socket and DDR2 RAM, with either clock speed improvements and/or architectural changes...

Will Intel have an answer for Hypertransport? How will their implimentation of 64bits compare? (Were either/both of these benchmarks 64bit, or were they differently compiled?) I know that AMD has already doubled the registers from 8 to 16 -- what did Intel do in Conroe? What about memory latency? Has Intel put the memory controller on the CPU?

TDP number is 65 watts, give or take - that's not going to change between now and launch, the chip is done.

Hypertransport is a non-issue on single and most dual core machines (as evidenced by benchmarks). I'll take the faster machine with the machine over HTT any day.

Intel's 64-bit implementation is identical to AMD's.

NeilBlanchard wrote:Greetings,

Was it an FX60 (i.e. dual core), or was it an FX57; which IIANM is single core?

It is said in the anandtech article. FX-60 overclocked to 2.8Ghz.

A small but important point: we do not know what the power consumption of Conroe is, and anyway, that will probably change from what it is now, to when the production chips are sold.

It is supposedly much more competitive then the current models. Intel says up to 40% improvement. And I have seen them mention TDP number of 65 watts. I'm fine with that until we get some real test results.

And, we do not know what improvements AMD will make, with the AM2 socket and DDR2 RAM, with either clock speed improvements and/or architectural changes...

Yes, we do know. AMD has announced the changes and they can be seen on their latest roadmaps. But we don't know how it will perform. The only AM2 benchmark to date is by THG where they used a prototype processor that supposedly has some problems. It was around 5% slower then a processor with DDR memory and same clock frequency. Not hardly a test that will show the true effect of DDR2.

Will Intel have an answer for Hypertransport?

Nothing as good as HT, but if they get more fps I'm satisfied even if they have inferior architechture.

How will their implimentation of 64bits compare?

Probably the same way, like they have done now.

(Were either/both of these benchmarks 64bit, or were they differently compiled?)

Had you read the anandtech article you could have seen from one of the pictures that it was not 64-bit version.

I know that AMD has already doubled the registers from 8 to 16 -- what did Intel do in Conroe?

Lot's of things. Too many for me to list them here.

What about memory latency? Has Intel put the memory controller on the CPU?

Integrated memory controller isn't on that list. And thus Intel clearly has bigger memory latency, although I don't mind the slower memory latency as long as it is faster...

jvrobert wrote:

NeilBlanchard wrote:Greetings,

Was it an FX60 (i.e. dual core), or was it an FX57; which IIANM is single core?

A small but important point: we do not know what the power consumption of Conroe is, and anyway, that will probably change from what it is now, to when the production chips are sold. And, we do not know what improvements AMD will make, with the AM2 socket and DDR2 RAM, with either clock speed improvements and/or architectural changes...

Will Intel have an answer for Hypertransport? How will their implementation of 64bits compare? (Were either/both of these benchmarks 64bit, or were they differently compiled?) I know that AMD has already doubled the registers from 8 to 16 -- what did Intel do in Conroe? What about memory latency? Has Intel put the memory controller on the CPU?

TDP number is 65 watts, give or take - that's not going to change between now and launch, the chip is done.

Hypertransport is a non-issue on single and most dual core machines (as evidenced by benchmarks). I'll take the faster machine with the machine over HTT any day.

Intel's 64-bit implementation is identical to AMD's.

If the benchmarks were comparing an FX57 with Conroe, then they were unfair.

Hypertransport is quite important with dual cores, and only testing with a multithreaded program, or with two programs at once, will show the difference.

If the real maximum wattage is 65watt, and if the idle is low enough, then that would be impressive -- and we will finally have a real competition on our hands. Because, that's around where the X2's already are, and AMD will be moving to .65u with AM2, right? We'll see where the performance/watt comparison comes out, when the Conroe comes out -- which is in about 6 months, right?

Streaming a gaming benchmark, or a media benchmark only tells us about bandwidth, not latency. Again, with the AM2, Athlon 64's with get DDR2 which will likely be a bigger boost for them, than it was for Intel.

Intel had not implemented the extra 8 registers in the P4 and Xeon, AFAIK. They had just essentially done the additional memory addressing? Hence their name "EM64T" -- Extended Memory 64 Technology", or something like that.

I'm not sure how important it is who beats who - consumers (and silent PCers) can only benefit whatever chip is best. At least AMD forced Intel to up it's game and if Intel takes the performance per watt lead, I'm sure the opposite will occur. It's a win win situation either way.

Is there any news on EIST/ CnQ equivalent for Intel's new desktop chips?

accord1999 wrote:

rpsgc wrote:This has been discussed on another thread. And how do you know its a FX-60 ? They only provide you with a POST screenshot showing "Unknown AMD Processor". And in the graphs they typed "Athlon 64 X2 @ 2.8GHz". So who's is it? FX-60, X2 @ 2.8GHz or Unknown ?

Intel did provide the testbed and benchmarks, I'll wait for something less "controversial".

The scores are essentially the same for FEAR as compared to a 2.8GHz FX-57 with the same video card setup.

http://www.anandtech.com/video/showdoc.aspx?i=2679&p=12

BINGO.

I noticed that as well.

These are competing in this test, but people have forgotten that such results have ALREADY been seen on a fx-57 !!

I am glad to see conroe is doing well. i mean, no im not.

I hate intel.

lets just not buy their crap.

tdp numbers for intel do not reflect things like actual wattage usage and real life experience.

um. yeah basically, they are useless for the most part. plus, intels drive the motherboard to use noticably more power. shrugs.

I will have to see a conroe going for a few days in a 75 degree F room with silent cooling to really believe intel hasnt made another hot chip.

Oh you say Dothan?

How about reading back on dothan once it was overclocked to gaming speeds of 2.5-2.6. yup, it got quite hot, wasnt very mobile anymore.

The very fact that intel did not put out wattage draws at the show is proof of something ugly. it wouldnt be hard to get two 40 dollar power angels and hook em up... it would be hard though to let everyone know that the chips arent cool running at those fps's.

mattthemuppet wrote:I'm not sure how important it is who beats who - consumers (and silent PCers) can only benefit whatever chip is best. At least AMD forced Intel to up it's game and if Intel takes the performance per watt lead, I'm sure the opposite will occur. It's a win win situation either way.

Bingo!

The technological edge held by any company will come and go in phases like any other business. The key here is competition. If AMD was not around, does anyone sincerely think that Intel would have abandoned their lame ass NetBurst direction so soon and switched entirely to their Israeli processor designs?

With Intel reawakened with real innovation instead recycling - hmm since all these new cores are based on the original Pentium's I guess thats not a valid statement. Now its AMD's turn to see if they can move beyond their aged Athlon designs - remember, the Athon 64's are still rooted on the older K5 and K6 designs.


I use whoever has the processor technology edge when I need new boxes:
I used Intel for the 8086 to the 80286 days.
Then to AMD for their 80386's.
I switched back to Intel for their Pentium I-III.
Then back to AMD for the XP Mobiles and the A64 Venice's and DC Opetrons.
My next upgrade just may returning to Intel for the Conroe class processors.

Let's hope it doesnt take AMD aquiring another NexGen like company in order to leap foward again.

NeilBlanchard wrote:
Hypertransport is quite important with dual cores, and only testing with a multithreaded program, or with two programs at once, will show the difference.

Intel has the superior system this time around, with the shared L2 cache and direct L1-L1 communication.

Intel had not implemented the extra 8 registers in the P4 and Xeon, AFAIK. They had just essentially done the additional memory addressing? Hence their name "EM64T" -- Extended Memory 64 Technology", or something like that.

It did implement them, EMT64 is functionally identical.

~El~Jefe~ wrote:tdp numbers for intel do not reflect things like actual wattage usage and real life experience.

um. yeah basically, they are useless for the most part. plus, intels drive the motherboard to use noticably more power. shrugs.

I will have to see a conroe going for a few days in a 75 degree F room with silent cooling to really believe intel hasnt made another hot chip.

Oh you say Dothan?

How about reading back on dothan once it was overclocked to gaming speeds of 2.5-2.6. yup, it got quite hot, wasnt very mobile anymore.

Which ones are those?

http://www.techreport.com/reviews/2005q ... ex.x?pg=17
http://www.gamepc.com/labs/view_content ... 780&page=3

Still more mobile than a Turion ML.

The very fact that intel did not put out wattage draws at the show is proof of something ugly. it wouldnt be hard to get two 40 dollar power angels and hook em up... it would be hard though to let everyone know that the chips arent cool running at those fps's.

Not quite the same thing, but they did show the power draw of a dual socket, dual core 3GHz Woodcrest server vs dual socket, dual core 2.4GHz Opterons server

http://www.xbitlabs.com/articles/editor ... 6-2_2.html

mattthemuppet wrote:I'm not sure how important it is who beats who - consumers (and silent PCers) can only benefit whatever chip is best. At least AMD forced Intel to up it's game and if Intel takes the performance per watt lead, I'm sure the opposite will occur. It's a win win situation either way.

Is there any news on EIST/ CnQ equivalent for Intel's new desktop chips?

I'd assume they're going to implement some sort of dynamic frequency and voltage adjustment based on load, but I don't have anything to back this up.

I did, however, read on Xbit Labs that all lines of CPUs (mobile, desktop, and server) will support advanced power gating. Isn't the technology that allows, for example, one core to be shut down during non-SMP tasks? If so, then this is a good thing for silencers I'd say.

Here's the article

And I don't think 65w is that hard to believe for Conroe. Pentium 4 didn't run hot because Intel sucks at making processors, it ran hot because it was a horribly inefficient design. Conroe, however, like other short-pipe CPUs (Athlon64, Pentium M) should run very cool, especially because it is @ 65nm.

But keep in mind TDP for the Conroe XE is quite a bit higher, 95w I believe. Still lower than even the most efficient top-of-the-line CPUs, though -- FX-60 was measured at around 110w by Xbit Labs, so around 95w or so for the CPU assuming the VRM is 85% efficient.

I'd assume they're going to implement some sort of dynamic frequency and voltage adjustment based on load, but I don't have anything to back this up.

wiki

The Intel Core Microarchitecture is designed from the ground up, but similar to the Pentium M in design philosophy. The data pipeline is 14 stages; a far cry from the 31 of Prescott. The cores will be 4-issue wide, instead of the 3-issue cores of P6, P6-M (Banias and derivatives), and NetBurst microarchitectures. The new architecture is a dual core design with linked L1 cache and shared L2 cache engineered for maximum performance per watt.

One new technology included in the design is Macro-Ops Fusion, which combines two X86 instructions into a single microinstruction. For example, a common code sequence like a compare followed by a conditional jump would become a single micro-op. Other new technologies include 1 cycle throughput (2 cycles previously) of all 128-bit SSE instructions and a new power saving design; the default stage for all components is minimum speed, allowing the chip to produce a minimum of heat, and consume as little power as possible.

Intel says that the power consumption of these new processors is to be extremely low; average use energy consumption is to be in the 1-2 watt range in ultra low voltage variants, with TDPs of 65 watts for Conroe and 80 watts for Woodcrest. However, this is subject to change. Merom, the mobile variant, is listed at 35 watts TDP for standard versions and 5 watts TDP for Ultra Low Voltage (ULV) versions.