Intel's Pentium 4 600 series

[DAPUNISHER]

Originally posted by: Zebo

Originally posted by: DAPUNISHER
outside of gaming the 600 series is as good a choice as any right now. .

Guess we don't read the same benchmarks. the lowly 3400 , which $190 BTW, put up a valiant fight to intels highest end chip the 3.76 EE costing hundreds (a thousand?) more.
http://www.anandtech.com/printarticle.aspx?i=2353
In anands review the FX's and 4000's simply dominate. Add to that lower power and AMD already has all the features intel recently incorporated only a dupe or rabid fanboy buys intels.

Hadn't seen Anand's benchies, but most corporate types are dupes by definition Features will=Intel wins the sale.

 

[Looney]

Originally posted by: LTC8K6
Hardcore, you mean you don't want 64 bits?

That's the question... do i want to pay $150 more simply for 64bit? Especially since the performance of 64bit just isn't up to par with AMD? The lower heat is a nice surprise, but it's not that important to me, even in a laptop. But looks like i have no choice, seeing as there isn't an AMD with a X800 or 6800Go coming out any time soon.

But talk about disappointing. Double the cache does almost nothing??

 

[PetNorth]

Originally posted by: LTC8K6
Opteron E power

I asked about some real test, not about nonsense Ed articles.

Do you know AMD's TDP is max for a whole family? Have you thought 0.9nm Opteron E TDP includes first dual core? Yes, it seems:

http://www.xbitlabs.com/news/cpu/display/20041230151732.html

Advanced Micro Devices? dual-core server processors to be released in mid-2005 will consume from 30W to 95W of power and will operate at core-speeds of 1.60GHz to 2.40GHz, according to data presumably sent by AMD to its partners among server components makers.

 

[LTC8K6]

I'll trust AMD to know the correct info.

Since AMD gives the Opteron 242 as 85.3 watts, and the 252 as 92.6 watts, and these numbers are much higher than earlier 90nm Opterons, I'll take their word for it that they are a lot hotter than the earlier 90nm Opterons.

 

[VirtualLarry]

Originally posted by: LTC8K6
Carlos, are you the only one who doesn't know that AMD and Intel have cross licensing agreements and willingly share and use each other's technology? They copy from each other by agreement.
Where is your outrage over AMD "copying" SSE2 & 3?
HT still isn't useless, btw.

I was under the understanding that Intel's SSE3, and AMD's AMD64 ISA, weren't covered by those more general cross-licensing agreements, and that AMD had to license Intel's SSE3 specifically, in order to include it into their new CPUs. Is that not true?

 

[PetNorth]

Originally posted by: LTC8K6
I'll trust AMD to know the correct info.

Since AMD gives the Opteron 242 as 85.3 watts, and the 252 as 92.6 watts, and these numbers are much higher than earlier 90nm Opterons, I'll take their word for it that they are a lot hotter than the earlier 90nm Opterons.

They would be it if they wouldn't include dual core, but they probably include them. Notice similar TDP in xbitlabs link talking about dual core. I read (I don't remember where) actual Dual core prototypes are E step (they'll be launch in mid 2005). So, here you have the explanation. Or do you think is there some logical or hide and strange reason to believe E 0.09nm single core are a lot hotter than D 0.09m single core?? SSE3 perhaps? come on...

 

[coldpower27]

No it's because this is AMD second 90nm process is using Strained Silicon which increases clock frequency potential, but seems to do it at the expense of power consumption, and with 2.4GHz looking to be the max AMD's aiming for for Dual Core, that still leaves the question why the increase TDP to 92.6W on the 2.6GHZ Opteron's.

Though we can make this anlaysis, once we get our hands on the E0 stepping Venice and San Diego cores. Right now we don't have enough information yet if the new 90nm Opterons are hotter then the old ones with just 90nm SOI, instead of this new 90nm SOI DSL.

 

[LTC8K6]

PetNorth, I really don't see how those numbers can include dual core at all. There's not enough wattage to account for a dual core, imo.

 

[slash196]

I think one equation says it all:

Intel = teh phail.

Man, AMD is looking pretty smart right now.

 

[clarkey01]

Originally posted by: slash196
I think one equation says it all:

Intel = teh phail.

Man, AMD is looking pretty smart right now.

Not really, its always a mixed bag

 

[clarkey01]

Originally posted by: LTC8K6
PetNorth, I really don't see how those numbers can include dual core at all. There's not enough wattage to account for a dual core, imo.

Give AMD some credit, I believe those numbers to be correct, its whats been said by a few sources inc xbitlabs.

 

[coldpower27]

Does Intel really need the performance crown or power conusmption crown? I mean they are still making plenty of money without it, as bad as Prescott is in the thermals and performance department compared to Athlno 64, it still brings in the bacon for Intel.

 

[clarkey01]

Yep and sad thing is they'll always win despite having a lesser design (not that thats always the case)

 

[PetNorth]

Originally posted by: coldpower27
No it's because this is AMD second 90nm process is using Strained Silicon which increases clock frequency potential, but seems to do it at the expense of power consumption, and with 2.4GHz looking to be the max AMD's aiming for for Dual Core, that still leaves the question why the increase TDP to 92.6W on the 2.6GHZ Opteron's.

Though we can make this anlaysis, once we get our hands on the E0 stepping Venice and San Diego cores. Right now we don't have enough information yet if the new 90nm Opterons are hotter then the old ones with just 90nm SOI, instead of this new 90nm SOI DSL.

Is out there some information about SS increases power consumition in these terms? (remember 25W more, this is a 37% increase) Do you thing it is reasonable? I think it isn't. I think, knowing what TDP is for AMD, and knowing what TDP first dual core will be (as Clarkey says, there are some believable sources about it), I'm sure first dual core are included in these E TDP.

 

[clarkey01]

Originally posted by: PetNorth

Originally posted by: coldpower27
No it's because this is AMD second 90nm process is using Strained Silicon which increases clock frequency potential, but seems to do it at the expense of power consumption, and with 2.4GHz looking to be the max AMD's aiming for for Dual Core, that still leaves the question why the increase TDP to 92.6W on the 2.6GHZ Opteron's.

Though we can make this anlaysis, once we get our hands on the E0 stepping Venice and San Diego cores. Right now we don't have enough information yet if the new 90nm Opterons are hotter then the old ones with just 90nm SOI, instead of this new 90nm SOI DSL.

Is out there some information about SS increases power consumition in these terms? (remember 25W more, this is a 37% increase) Do you thing it is reasonable? I think it isn't. I think, knowing what TDP is for AMD, and knowing what TDP first dual core will be (as Clarkey says, there are some believable sources about it), I'm sure first dual core are included in these E TDP.

37% increase for a 200 Mhz bump, and lol, we'r wanting 3 Ghz athlons ! I doubt we'll see them without Prescott TDP's or over.

 

[devildog45co]

I don't know if anyone is still talking about it on the thread, but I went over to OCAU to check oput the thread that Duvie was talking about. That Chainbolt guy should be the intel poster child. Or a pretzel company poster child. His logic is certainly all twisted up like one.

 

[coldpower27]

90nm SOI 67W 2.2GHZ
90nm SOI DSL 85.3W 2.4GHZ
90nm SOI DSL 92.6W 2.6GHZ

Your talking about 25W difference between 2.2GHZ and 2.6GHZ, so it's a 400MHZ increase. This is entirely possible if you need to account for a 400MHZ increase and the addition of strained silicon.

If we imagine it takes about 6W to go from 2.2GHZ to 2.4GHZ on SOI DSL That would put a 2.2GHZ 90nm SOI DSL part at around ~ 79W, which represents about an 18% increase in power consumption for the use of Strain Silicon which isn't that bad.

If AMD's 90nm SOI DSL parts are indeed hotter, then their 90nm SOI parts then it does show that it's not entirely Prescott's fault that it in itself runs so toasty.

Thought this would make the Dothan core even more amazing

Though in the end it would be fun to bench Venice beside the Winchester core to see how they compare at different clockspeeds.

 

[ribbon13]

http://amdzone.net/pics/cpus/dualcore/1stdemo/opt13090dualcore.jpg

 

[MemoryInAGarden]

A 2.4C@3.6 ghz w/ 2-2-2-5 BH-5 is still pretty much going to kick anythign Intel has put out since then. This is Feb.2005, that combo could have been had in the summer of 2003. That really goes to show how badly Intel is running in a circle.

 

[LTC8K6]

How long has A64 been out and what would the early ones overclock to? Pretty much the same story.

 

[1stoff]

my 2.8c @ 3.8 beats the new expensive edition 3.76 in Sandra CPU Arithmetic

my 3.8c: ALU 11791 MIPS........FPU/iSSE2 4798/8568 MFLOPS
3.76 EE : " 10765 " " " 4425/7656 "

 

[carlosd]

Originally posted by: LTC8K6
Carlos, are you the only one who doesn't know that AMD and Intel have cross licensing agreements and willingly share and use each other's technology? They copy from each other by agreement.

Where is your outrage over AMD "copying" SSE2 & 3?

HT still isn't useless, btw.

At least , AMD is not pricing their CPUs so high only for adding intel features.

And if HT is not useless so why in multitasking tetst made by anandtech P4 CPUs are still beaten by A64 Cpus wich doesn't have HT. Will the P4 perform worse with HT disabled. ohh!! so poor, so HT is needed to barely compete against A64 CPU`S ohh so useless.

 

[carlosd]

q]Originally posted by: 1stoff
my 2.8c @ 3.8 beats the new expensive edition 3.76 in Sandra CPU Arithmetic

my 3.8c: ALU 11791 MIPS........FPU/iSSE2 4798/8568 MFLOPS
3.76 EE : " 10765 " " " 4425/7656 "
[/quote]


Man, this a sinthethyc benchmarks , it is useless no real world performance reflect, try some real benchies.

 

[Duvie]

Originally posted by: MemoryInAGarden
A 2.4C@3.6 ghz w/ 2-2-2-5 BH-5 is still pretty much going to kick anythign Intel has put out since then. This is Feb.2005, that combo could have been had in the summer of 2003. That really goes to show how badly Intel is running in a circle.

That was rare to see a 2.4c do 3.6ghz...i know cause I had 3.5ghz back in Feb of 2004 and that was rare...It wasn't until the MO stepping chips arrived in 2Q of 04 that we saw some 3.6ghz out of them...

With that I agree the 3.6ghz would beat almost all chips Intel has been able to muster since...sad..truly!!!

 

[christoph83]

Originally posted by: carlosd

Originally posted by: LTC8K6
Carlos, are you the only one who doesn't know that AMD and Intel have cross licensing agreements and willingly share and use each other's technology? They copy from each other by agreement.

Where is your outrage over AMD "copying" SSE2 & 3?

HT still isn't useless, btw.

At least , AMD is not pricing their CPUs so high only for adding intel features.

And if HT is not useless so why in multitasking tetst made by anandtech P4 CPUs are still beaten by A64 Cpus wich doesn't have HT. Will the P4 perform worse with HT disabled. ohh!! so poor, so HT is needed to barely compete against A64 CPU`S ohh so useless.

You make it sound like intel's chip loses every test, when it won 3 out of 5. Quite honestly I don't understand all the negativity in general. For gamers and performance nuts like us AMD is a clear choice right now. But for certain apps, intel still dominates.

Instead of looking at all the green bars near the top, check out the chips at the bottom. The 630 beats the 3000 a64 in EVERY audio/video encoding test, and even beats the 3400 and 3200 in some.

The 630 wins in photoshop over the 3000 and beats the 3200 in roxio. The 630 is above the 3200 in EVERY multitasking test except the first one where its within 3%.

In the General usage tests the 630 beats the 3200 3 out of 10 times, it beats the 3000 in 2 and is within 5% of the 3000 on 2 tests.

Im Rendering the 630 beats the 3200 1 out of 3 and the 3000 on another.

In workstation the 630 beats the 3000 5 out of 8 and beats the 3400 sometimes.

And where the chip is most obviously lacking, gaming, the 630 stays within 5-10% of the 3000.

I know the 630 costs more but im betting we'll see a price drop, and the 530 is pretty similiar to the 630. For 30 bucks more than the 3000. Now for gaming, its clear go amd, but for 30 bucks extra, the 630/530 are winning many other tests AND most importantly, in some that it loses, keeping close. To say this chip sucks is really uncalled for. Now on the otherhand, their EE offerings and higher end chips, those are not worth it, just like going after an FX or even 3800.

People act like the general public is getting ripped off by intel when most of their volume in desktops is more than likely in the 2.8ghz - 3.2ghz range, where these chips, according to anandtech's test, hold up a lot better than their high end offerings.

The general consensus here seems to be that these chips suck, while in gaming it does lag behind, in everything else...it does NOT. And I'm not trying to say these chips rule and you should go buy intel now. I'm just giving a well needed different perspective.