real conroe test by victorwang

[iterations]

About a week ago (maybe even a little longer) I think I saw some mention in the THG forumz that Intel had sent out engineering samples of Conroe to several people. Then the intervening silence made me start wondering if I was dreaming. Like, where are the benchmarks. Now I know I wasn't dreaming. But I am wondering when the rest of the lucky ES receivers will post benchies? I thought some established webzine sites like anandtech or even THG were included. Wish I could find those posts again. The posts by victorwang also made a few things make more sense. His apparent struggles to get the setup running, and the general flakiness, explains to me why Intel only allowed "their" benchmarks to be run at IDF. Wouldn't look good if they allowed unrestricted benchmark installations, and a bunch of them crashed the system at such a high profile venue. This also explains why Intel hasn't already released the Conroe. The platform still needs some ironing out, and not because of some conspiracy.

Just so you understand, the "proper" sites that you mention are all still currently under NDA. Victor got his conroe from ebay and therefore was not bound by an NDA (nor was he supplied a working motherboard, etc...) so he has pieced something together as best as he could, given that mainstream parts are not generally available yet. Even then his benchmarking results are great, and most welcome in the community.

Platform support for Conroe should not be a problem when it actually gets released. Because it uses signifcantly less power than the Prescott Pentium D's it is replacing, the motherboards needs a voltage regulator change so they can be tuned for the smaller energy needs of Conroe. In some cases (Such as Intel's BadAxe 975 board) this will be as simple as a firmware flash.

Once the NDA is up you will see a plethora of reviews from all your favorite sites.

Cheers!

 

[clarkclark]

I think the hard drive lags the system most right now. HD is too slow to compare with other elements in the system. The Ram Disk from Gigabyte is a good example that solid state storage system will replace the old HD technology. But, now, the price for the ram is very high in unit of $/M. If the price of DDR RAM drops or even the old SDRAM, we can definitely replace the HD.(even 100Mhz SDRAM has a bandwidth of 800M/s and that's way faster than HD). The Ram Disk from Gigabyte has 4 slots for DDR. It is possible to make 64 slots on both of the card.(64G) and the if we install 2 of them in the PCI slots, the total space is 128G. That is really sweet..the only negative factor is that the solid state storage needs a voltage to keep the information..therefore a big rechargeable battery is necessary for the storage. ..

What is inside the RAM? well, the product from sand...and a bunch of internal wires

 

[ryokinshin]

i dont know why these people are getting all excited, when the retail product comes out, we will see the real benchmarks

 

[clarkclark]

because they are excited from valance band to conduction band 8O

 

[WINDSHEAR]

I doubt they will do that. They always want to make money, they don't want to save us money. As a consequence, I don't think we'll get to use our old pc100 sdram sticks

it'd sure be nice though. I'd like to lobby for it.

 

[fainis]

i also hope it has a good performance per price ratio....

i can hardly wait to see the prices

any of you guys heard anything about temperatures and power consumption, TDP samething?


obviously that 65nm should make a difference.....

90nm is good for single core...we have all seen what makes on venice and 6xx series on intel......good temperatures and low power consumption...i bet 65 will be much better ...... i wonder ...... will it be enough for 4 cores or even more.........45nm is there...hyhyhy

 

[iterations]

i also hope it has a good performance per price ratio....

i can hardly wait to see the prices

any of you guys heard anything about temperatures and power consumption, TDP samething?


obviously that 65nm should make a difference.....

90nm is good for single core...we have all seen what makes on venice and 6xx series on intel......good temperatures and low power consumption...i bet 65 will be much better ...... i wonder ...... will it be enough for 4 cores or even more.........45nm is there...hyhyhy

Based on the roadmap prices, it looks to be very good. The 2.4Ghz Conroe, which beats even an AMD FX-60 ($1200) on almost all benchmarks done so far, will be priced at $309 at introduction. A 2.67GHz version looks to be $527, and there is also a $240 and $210 version as well, but can't remember their clockspeed at the moment.

Power consumption also looks to be incredibly low. All of the Conroe chips are to be within a 65W or lower power envelope (TDP), except for the 3.33GHz extreme edition with looks to come in at around 83W. This is a HUGE reduction is energy usage compared to Intel previous desktop offerings, and even AMD's products.

The only temperature data we have so far is from the leaked chip that Victor Wang got. It appeared to hold at 26C when fully loaded. He didn't even need the heatsink fan. Sounds pretty good so far.

 

[tamalsmith]

i also hope it has a good performance per price ratio....

i can hardly wait to see the prices

any of you guys heard anything about temperatures and power consumption, TDP samething?


obviously that 65nm should make a difference.....

90nm is good for single core...we have all seen what makes on venice and 6xx series on intel......good temperatures and low power consumption...i bet 65 will be much better ...... i wonder ...... will it be enough for 4 cores or even more.........45nm is there...hyhyhy

Based on the roadmap prices, it looks to be very good. The 2.4Ghz Conroe, which beats even an AMD FX-60 ($1200) on almost all benchmarks done so far, will be priced at $309 at introduction. A 2.67GHz version looks to be $527, and there is also a $240 and $210 version as well, but can't remember their clockspeed at the moment.

Power consumption also looks to be incredibly low. All of the Conroe chips are to be within a 65W or lower power envelope (TDP), except for the 3.33GHz extreme edition with looks to come in at around 83W. This is a HUGE reduction is energy usage compared to Intel previous desktop offerings, and even AMD's products.

The only temperature data we have so far is from the leaked chip that Victor Wang got. It appeared to hold at 26C when fully loaded. He didn't even need the heatsink fan. Sounds pretty good so far.

dont get cooky.

63 watts AM2 will be once AMD switches to 65mm, by the time these 65W envelope comes out for intel ( 45mm )
so far AMD still seems to win in eficiency.
but atm conroe seems clearly superior in speed.
let's hope AMD gets a boost with that new Z-ram with 45mm Cache they were talking about in anantech with IBM & other company

 

[iterations]

dont get cooky.

63 watts AM2 will be once AMD switches to 65mm, by the time these 65W envelope comes out for intel ( 45mm )
so far AMD still seems to win in eficiency.
but atm conroe seems clearly superior in speed.
let's hope AMD gets a boost with that new Z-ram with 45mm Cache they were talking about in anantech with IBM & other company

I think you are slightly confused. The 65W TDPs for the Conroe chips are for the current 65nm process that they will debut at in the next couple months. This is a fact.

The 63W that you claim for AMD is indeed for the _lower_power_ version of their AM2 chips _when_ they switch to 65nm, which should hopefully happen by the begining of 2007. As for zRam, it is just a technology that they are looking at and had a press release about. It is not even on their roadmaps for products yet. Also if you read the zRam whitepapers your see that while it is small and low power, it has far to large of a latency to be used as an L2, and is much more suitable for an L3 cache. Even so zRam looks very interesting as a future technology.

Just wanted to clear up this discussion with a few facts. It is intellectually dishonest (or simply misinformed) of you make the claims you just did.

Also, you should note that a TDP is merely an upper limit, the actual amount of energy will vary from 0W (off) to ~3W when idle in deep sleep state to ~65W in the worst case when fully loaded.

Cheers!

 

[xSunnYx]

For those who intrested...

http://tweakers.net/aanbod/advertentie/35855

http://wausmod.computerjunkie.nl/results/17.JPG

 

[fainis]

what`s that my friend? .. that`s a yonah

 

[clarkclark]

3G yonah=2.4G conroe

 

[Atolsammeek]

I think people need to still wait and see. Intel Fanboys need to pull there panties out of there rears also Amd need to wait and see. 6 weeks is not too far off.

 

[Grimmy]

Trouble is... there will ALWAYS be a next time.

Its been said time and time and time again. Leap Froggin :lol:

 

[RichPLS]

au contraire... uno instance will be finale...

 

[clarkclark]

the Conroe+965+ICH8 will be landing in june

 

[Parrot]

Real Conroe test will only come out when the REAL CONROE comes out. Maybe this is right, maybe it's BS--time alone will tell, not theinquirer.net .

Word.
Have you checked out Victor Wang's profile? Well I have and amongst other things under "Biography" it says"Intel". Does this mean he works for Intel?
I smell something fishy!!!!!

 

[gOJDO]

Real Conroe test will only come out when the REAL CONROE comes out. Maybe this is right, maybe it's BS--time alone will tell, not theinquirer.net .

Word.
Have you checked out Victor Wang's profile? Well I have and amongst other things under "Biography" it says"Intel". Does this mean he works for Intel?
I smell something fishy!!!!!
I would like to listen and learn from someone who have working expirience in any company like Intel, IBM, AMD, Motorola and etc. And I would like not to listen becouse there is nothing to lear from fanboys, no metter if they are AMD or Intel fans.
Can you tell me where I can find more about Victor Wang?

 

[Parrot]

Real Conroe test will only come out when the REAL CONROE comes out. Maybe this is right, maybe it's BS--time alone will tell, not theinquirer.net .

Word.
Have you checked out Victor Wang's profile? Well I have and amongst other things under "Biography" it says"Intel". Does this mean he works for Intel?
I smell something fishy!!!!!
I would like to listen and learn from someone who have working expirience in any company like Intel, IBM, AMD, Motorola and etc. And I would like not to listen becouse there is nothing to lear from fanboys, no metter if they are AMD or Intel fans.
Can you tell me where I can find more about Victor Wang?

Ok Intel fanboy, what are you getting so excited about??If he works for Intel, then you should be happy-right??I am just trying to get at the truth. If someone posts benchmarks for an Intel product then I think we should know if he is an Intel employee. The same goes for ANY other company and its products!!

 

[MrsD]

NO CONROE TEST IS GOING TO BE LEGIT BECAUSE ITS NOT BEEN RELEASED YET!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
ENGINEERING SAMPLES DO NOT COUNT AS THEY ARE NOT FINAL PRODUCT THAT WILL BE AVAILABLE TO CONSUMERS.
GET A LIFE NERDS!

 

[RichPLS]

Shave you hairy sheeplet. :lol:

 

[cubicleslave]

NO CONROE TEST IS GOING TO BE LEGIT BECAUSE ITS NOT BEEN RELEASED YET!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
ENGINEERING SAMPLES DO NOT COUNT AS THEY ARE NOT FINAL PRODUCT THAT WILL BE AVAILABLE TO CONSUMERS.
GET A LIFE NERDS!

Hello all, I got a train of thought that I find very compelling, and hope to get some constructive feedback (please, no fanboys or flames). It makes sense to me when people claim that the conroe benchmarks to this point in time, may or may not be representative of the final released product's performance. This is because the product might change between now and the release date. However, I don't see what's wrong with using the current benchmark results as an indicator of said performance, because what we see now, if it is in error, would err on the side of conservatism. This is because whatever comes out at release time, if it is DIFFERENT, can only be better, not worse. After all, why would Intel's engineers do something to cripple the engineering sample version and then release it as the final product? I think they'd have better things to do with their time.

One possible argument that I have seen, is that the ES conroes could be cherry-picked high performance units. However, that concern doesn't really stand up to scrutiny. Cherry-picking will yield parts that are higher on the process skew distribution bell-curve (and is probably how the Extreme Editions are obtained). In other words, they can run at higher clock speeds. Logically and architecturally, all Conroes that come off a fab run are identical. They only vary in their speed as far as performance is concerned. You will not be able to cherry-pick some units that are 3-issue Conroes and some that are 4-issue, or some with shared level2 cache and some that have independent level 2 cache. Remember, a cpu is a deterministic finite-state machine. This means that if a released part runs at 2.66GHz, then it should yield the same performance as the engineering sample that ran at 2.66GHz, unless Intel's engineers changed something in the released version. But if in fact they did change something, the change can only improve, not hurt the performance (see my previous comment about engineers having nothing better to do with their time).

Does this all make sense? In conclusion, I think we should enjoy and appreciate the efforts of FCG, VictorWang, FUGGER, etc for their generosity with their time, and also recognize that if the final product comes out different, it can only be better and not worse. It cannot be othewise, unless an Intel engineer somewhere made a career-ending idiotic mistake. But even if that were the case, Intel could still fix the error and eventually release a part that is as good or better than the ES version. If there is a flaw in my reasoning, please point it out and explain why it is flawed. Again, please, no flames or fanboys! Thanks a lot!

 

[Action_Man]

And the trolling campaign continues.

 

[endyen]

I guess we have different views of what a troll is.
For you, it would seem that anyone who opposes Intel is a troll.
From where I sit, conroe will retake the crown. But only by a small margin.
If they hit the stores at advertized prices, they will be good value.
If, on the other hand, conroe gets it's butt kicked on non-floptimized progs, or if the price guestimates are low, Intel may not be worthwhile still.
Does that make me a troll too?

 

[Action_Man]

I guess we have different views of what a troll is.

Obviously.

For you, it would seem that anyone who opposes Intel is a troll.

Inncorrect. Stop labelling me an Intel fanboy, its insulting for both of us.

If they hit the stores at advertized prices, they will be good value.

Why wouldn't they hit the shelves at advertised price?

If, on the other hand, conroe gets it's butt kicked on non-floptimized progs

I didn't know the games they used were floptimised.

Does that make me a troll too?

Nope.

If you'd have read most of parrots post you'd know he is a troll and it is a campaign. Why bump an old thread when hes already brought it up? :roll:

Besides his point isn't even really valid. Other people have run tests as well and I'm quite certain they're not Intel employees. I also doubt victor wang is as well.