Will AMD regain OC crown off Pentium D 805, is AMD slipping?

[desilver]

A short while ago we tested Intel's latest flagship processor, the Pentium EE 965 (Extreme Edition), which costs nearly $1,100 at retail outlets. Even this CPU, which still isn't available at too many locations, has to surrender first place to this stealth candidate. Things look the same for the top-of-the-line AMD processor, the Athlon 64 FX-60, which also fell behind in most of our benchmarking categories.

http://www.tomshardware.com/2006/05/10/dual_41_ghz_cores/page7.html
A shrewd move by Intel, it has the hall marks of an Intel come-back.
The fact that Intel has won the OC crown back, has 45nm wafers set to go Conroe in the wings,and AMD is stuck on 90nm---with just a new name for its AM2 platform, with renamed processors,does this spell doom and gloom for AMD? Will it survive the $M + Intel joint venture, or will it end up dead and buried?

 

[powerbaselx]

A shrewd move by Intel, it has the hall marks of an Intel come-back.
The fact that Intel has won the OC crown back, has 45nm wafers set to go Conroe in the wings,and AMD is stuck on 90nm---with just a new name for its AM2 platform, with renamed processors,does this spell doom and gloom for AMD? Will it survive the $M + Intel joint venture, or will it end up dead and buried?

Don't be so pessimist. This is good for the consumer: more flexibility, more improvements, more performance, less power consuption, for less money!

This is like the mouse and the cat. Hope it never ends.

 

[S7A88Y]

WTF are you talking about...GOTH. AMD will have 65nm running smoothly on AM2 and Intel has sampled unstable 45nm chips. Also people forget AMD has IBM's R&D team. They have made carbon nanotube ICs up to 50Mhz. By the time Intel goes to a STABLE 45nm so will AMD or AMD will be ready for carbon nanotubes, allowing a giant leap in preformance, better than 15nm. So do research before talking about such a topic.

Let me add in that I KNOW 65nm Athlons will overclock very well.

 

[shabodah]

I want to know how you know. I mean, I suspect it, myself. The 90nm athlons OC far better than the 130's did, but I don't know where you're getting any of this info. Send us some links. Another point to make is the fact that although the 805 offers great OC'ing potential, it sure does make a LOT of heat and take a LOT of power to get there. An Opty 170 can perform 95% as good on 25% of the power and giving off 25% of the heat.

 

[powerbaselx]

WTF are you talking about...GOTH.

Goth?! LOL!! :-D

Anyway, as you can see, there are people immediatly happy with the "new" intel 805 chip... i'm not saying that is better technically than the AMDs. What i'm saying is that some competition is good for the consumer.

 

[desilver]

As far as I am aware,
AMD 65nm is only in testing with AMD right now....
and Intel 45nm in testing with Intel right now....

...note, NOT in production = does NOT exist!! 8)

so what are you taking about...? :roll:

AMD are ...looking for a partner to produce it..(rumor only)...in Singapore.

 

[Heyyou27]

WTF are you talking about...GOTH. AMD will have 65nm running smoothly on AM2 and Intel has sampled unstable 45nm chips. Also people forget AMD has IBM's R&D team. They have made carbon nanotube ICs up to 50Mhz. By the time Intel goes to a STABLE 45nm so will AMD or AMD will be ready for carbon nanotubes, allowing a giant leap in preformance, better than 15nm. So do research before talking about such a topic.

Let me add in that I KNOW 65nm Athlons will overclock very well.

There's nothing but opinionated crap in your post. :roll:

 

[MG37221]

To my knowledge, AMD hasn't held an overclocking crown since the release of their 386DX-40MHz CPU. They've exchanged the performance crown with Intel several times since then but Intel procs generally overclock better than AMD, and for the most part, always have.

This isn't to say that they necessarily scale as well, especially with the current AMD CPU technology, where a 200MHz clock speed increase is a serious and noticable increase in performance. But when you look at highest clock as a percentage over stock, Intel has always held that crown.

 

[gomerpile]

I bought the processor I must say it clocked no problem to 3.1 and by far is a good all round processor, I will be doing more test but I bought it for my daughter and she will be so mad if I take it out for mine so tonight she has a sleep over and I'm switching from the xp3200 to see the benchmark difference. But I can tell you with what she has a cheap mb and ram and the cpu bench was by far passed the xp no problems

 

[Mike995]

Amd never had the overclocking crown, my amds dont overclock very well at all. How ever my p4 overclocked 1.2 ghz stock volts, and air. No problem, as if it was factory clocked at 3.6 ghz. But my amd system using the same power supply, ram, and video card struggles with a 400 mhz overclock. And im using a venice core. S788by your wrong... completly wrong, amds silicon process doesnt come close to intels, and its a fact that amd processors dont last as long while overclocked as intel processors. Amd doesnt even have 65 nm out yet, intel has for what... 8 months. 45 nm intels may be out the 1st quarter of next year. How would you know they are buggy. Intel can shrink sram cells smaller then amd or Ibm, on the same lithography size. It also laughably took amd 4 respins to get the new memory controler functioning properly on their AM2 processors, the conroes are overclocking 1 ghz-1.6 ghz on the first spins of the chip.

 

[theonejrs]

desilver,

That's a pretty lame claim for an Intel fan to make. First off, I just ordered a P-4D 940. Why, because I don't have to OC the hell out of it to catch up to the competition. If I want, I can OC it 10% - 15% with no problem and stomp the D805 in the process. Who wants to own a ticking time bomb anyway, not me! The heat issues alone at those speeds would drive me nuts. Tomshardware says that you need water cooling over 4.0. The temps I've seen posted for this CPU are alarming. I don't consider an idle temp of 60C and encoding at 69C very good at all. It won't do any of the other componants any good either! I don't have to OC the hell out of the D940 to make it fast and reliable. My guess is that all that heat will take it's toll! The down side of the D805 is if you slow it down to where the temps are reasonable, it's a dog by comparisum to other dual-cores!

Not only that, lets be fair and OC all the other CPUs in the test and watch how far the D805 drops down the list. It's rediculous to compare benchmarks with a highly OC'd CPU against stock ones! Put it up against an FX-60 or a 185 Opteron overclocked and see how well it does. I wish someone would do just that. Intel hasn't taken the speed crown yet. They will when the Conroe comes out. AMD has acknowleged that the Conroe will be about 20% faster. They were so calm about when the press conference was held, you know thay are ready for this. I can't wait to see what they have in store for us. It should be interesting!

Happy Computering

At present, I run a 3.0/800/HT Prescott at 3.60 (20% OC) with memory timings of 2 2 2 5 so I know a little about heat and overclocking. I have a Zalman 9500 LED which will work very nicely on the D940 as it came with the 775 mount.

 

[theonejrs]

gomerpile,

There isn't any single core CPUs that can out benchmark a D805 stock at 2.66. Not even the fabled FX-57 which is probably the fastes single core out there. In fact, no single core can come close to matching any dual-core in a CPU benchmark! My numbers from my 3.0/800 Prescott, running at 3.6 which runs about 11 thousand or so mips should jump to about 23 thousand mips with the D940, and that's stock at 3.2!

Happy Computering

 

[Ycon]

Never heard such cr*p before :lol:

 

[theonejrs]

ycon,

It may be crap to you but here's the benchmarks to prove it!

And here's the memory bandwidth!

And the temps while encoding!

I don't say something unless I can back it up!

Happy Computering,
theonejrs

 

[theonejrs]

sorry my error

 

[noblekitty]

The whole point of THG recent article about the D805 is to show that a budget dual core @ $130 can be overclocked and rival CPUs @ the $1000 price range.
If you can't get that simple point, stop bashing the article. It did not say anything about the D805 is the fastest CPU compare to the other Ds or the FXs.

About the heat and power thing, I am pretty sure that your D940 will run a little better than the D805 @ same clock speed all because it is a 65nm process, thats all. Most same process D cpus, when running at the same speed, should produce the same amount of heat or very close to.

The D805 is intentionally clocked slower at default by Intel, because of marketing strategy. It is very capable just like any other D8XX.

 

[theonejrs]

ycon,

Sorry bout that, I got the benches from before I went tp XMS mmeory. Here's the correct ones.

ycon,It may be crap to you but here's the benchmarks to prove it!

And here's the memory bandwidth!

And the temps while encoding!

I don't say something unless I can back it up!

Happy Computering,
theonejrs

 

[mjp1618]

WTF are you talking about...GOTH. AMD will have 65nm running smoothly on AM2 and Intel has sampled unstable 45nm chips. Also people forget AMD has IBM's R&D team. They have made carbon nanotube ICs up to 50Mhz. By the time Intel goes to a STABLE 45nm so will AMD or AMD will be ready for carbon nanotubes, allowing a giant leap in preformance, better than 15nm. So do research before talking about such a topic.

Let me add in that I KNOW 65nm Athlons will overclock very well.

What are you on? First of all, AMD doesn't own IBM's engineers. It's more like AMD doesn't have any R&D, so they just borrow IBM's R&D results.

What do you mean Intel sampled "unstable" 45nm chips? Care to elaborate? You mean, they started jumping up and down the moment they were supplied power?

Of course, IBM has a very good R&D division, and they have Carbon Nanotube projects. NOT AMD. And IBM is not going to simply develop Carbon Nanotubes and give it to AMD for making CPUs.

Carbon Nanotubes before Intel's 45nm? You might as well say AMD will have Neutrino-based Quantum Computing before Intel's 45nm...

One advice for you: join Sharikou's club. You two will make a great couple.

 

[theonejrs]

noblekitty,

"The D805 is intentionally clocked slower at default by Intel, because of marketing strategy. It is very capable just like any other D8XX."

This is one of the silliest things I've ever heard! Do you honestly think that Intel wouldn't have brought this CPU out as at least a 3.0 or 3.2 if that was the case? First off, at the time Intel could have really used a shot in the arm. Sales were dropping and they were losing market share. my guess is that Intel wasn't even aware of the OC abilities of the D805 until Tom's Hardware posted their article. If Intel did know, they were stupid and missed a huge marketing opportunity at a time when they needed it the most!

As far as being capable, temps vs CPU speed will be the deciding factor on it's capabilities. Get it up to where it competes with the better chips and heat will be very much the issue. Drop it down to where the temps are easier to manage and it's a dog compared to other Dual-Cores. Summer is coming as well! It's going to take a very cool room to run one of these babies at 4.0 reliably. I know of one guy who has a Vapochill/compressor and his runs great but If I have to spend that much money then I'll buy an AMD Opteron 185 and be way ahead of the game! Mind you, at 2.66, the D805 is still faster overall than a single core but it won't ba a match for any of the other Dual-Cores shown in the benches if they OC too!

Happy Computering,
theonejrs

 

[noblekitty]

One advice for u,
You need to read more about how Intel and AMD set speed on their CPUs. Alot of CPUs are set to run slower than they can, by both companies.

Enough said!

 

[DaBigHurt]

Comparing OC speeds to stock speeds is like saying , :"Well, if I was faster, ran cooler, had more power and had good ram I would beat you no doubt about it"

 

[noblekitty]

????
Still can't understand the point after reading the post several times...

 

[powerbaselx]

noblekitty,

"The D805 is intentionally clocked slower at default by Intel, because of marketing strategy. It is very capable just like any other D8XX."

This is one of the silliest things I've ever heard! Do you honestly think that Intel wouldn't have brought this CPU out as at least a 3.0 or 3.2 if that was the case? First off, at the time Intel could have really used a shot in the arm. Sales were dropping and they were losing market share.

I agree also this is a marketing strategy rather than a technical issue. Probably Intel wanted to grow faster its sales and reach some market segments with this cheap D series chip. As we can see, it's becoming popular, well accepted from the beginning (even by THG) despite its 2.66GHz (slow) frequency.
To sale a cheap cpu with 3.0GHz or higher with those new capabilities could compromise Intel current offering much more expensive. And of course this cpu bypasses the heat issue...

my guess is that Intel wasn't even aware of the OC abilities of the D805 until Tom's Hardware posted their article. If Intel did know, they were stupid and missed a huge marketing opportunity at a time when they needed it the most!

Honestly, i really doubt Intel pays much attention to TMG (or other popular sites) articles and OC experiments... their main source are their business results, feedback from is own sales force and probably some consulting market reports...

Just my 2 cents...

 

[DaBigHurt]

What I mean its that some ppl get caugh up in the hype that OC Pentiums would beat AMD's at stock speeds. Regarding the recent "4.1GHz" P.

 

[chunkymonster]

The whole point of THG recent article about the D805 is to show that a budget dual core @ $130 can be overclocked and rival CPUs @ the $1000 price range.
If you can't get that simple point, stop bashing the article. It did not say anything about the D805 is the fastest CPU compare to the other Ds or the FXs.

This is the most sensible post in this entire thread.

AMD is slipping no more than Intel is slipping.

Another BS thread...