Intel Nehalem Turbo-charges Radeon HD4850 Benchmark

[lemonadesoda]

Nehalem manages between 0% and 40% performance increase clock-for-clock compared to penryn. But with a power consumption cost of 10% for the whole system, which is driven by about 20% extra power for the CPU.

http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3326&p=7

So, Nehalem, TODAY, is not impressive at all. You can get the same performance per watt by just overclocking a penryn. However, I'm sure the figures will improve once they optimise mainboard, BIOS, memory channels, and final (non-engineering sample) CPUs will have low power requirements. Or rather, lets hope so, otherwise Nahelem is a flop, and NOTHING like the mammoth win when Intel moved to Core 2 architecture.

 

[Mattgal]

who did this review is a dumb person! such a high-end cpu with a peace of crap vga! why wouldn't they tested it with a gtx280 or GX2? that way there will be no bottleneck

 

[Deleted member 3]

who did this review is a dumb person! such a high-end cpu with a peace of crap vga! why wouldn't they tested it with a gtx280 or GX2? that way there will be no bottleneck

Well, while we're calling people dumb people I would read the text again. It's the CPU score they're mainly talking about, so the GPU doesn't matter much. Then again, not being dumb, you already knew that.

 

[Mattgal]

i already sad why. there s a bottleneck and a pc cannot perform its best. talking from my own experiance

 

[btarunr]

Peace of Crap: That which exists between the United States and North Korea.

^Doesn't fit in your sentence.

Looking at a pair of dual Xeons just a couple of hundred points behind this and the fact that even with dual-channel, this chip is expected to have insanely low memory latency owing to an IMC, this chip can do a lot better than this, only time will tell.

That bench is just for you to look at the CPU scores in 3D benches, the 'HD4850' in the headline is merely to show there's no latest NVIDIA GPU that alters CPU score. And that I believe this is a fast bench for HD4850 compared to other reviewers using mid/mid-high range CPUs for evaluating a HD4850, this bench is just a novelty for "yay we were the first to do it".

 

[Mattgal]

i dont like ATi + nvidia GTX280 or GTX280 sli or even TRI-SLI is LOADS better than that thing! atleast a HD4870 !

 

[btarunr]

Whether you use a GeForce 6200 TC or GeForce GTX 280, the CPU scores will come the same. That's what you're expected to look at.

 

[InnocentCriminal]

I think you're missing the point (that btarunr is point out) Mattgal.

 

[Mattgal]

about a cpu score yes. but i still say that its bottlenecked. (tough for benches i dont think it will make a big difference) but in fps ecc HELL YESS

 

[btarunr]

You have zero clue what you're talking about.

This is a bench merely to show you what impact a Nehalem derivative has on the bench, of what CPU score it churns out. Is that hard to understand? A VGA will never bottleneck a CPU benchmark, only the reverse is possible.

 

[InnocentCriminal]

I think you're expecting too much from such an early example of Nehalem and (especially) the 4850, don't forget it's not a high-end card, it's mid-range.

 

[btarunr]

I think you're expecting too much from such an early example of Nehalem.

No, he expects the CPU score to cross a gazzillion billion after using better graphics hardware.

 

[yogurt_21]

hmm in 06 it's only 300pts faster than my quad at 3GHZ. I was expecting much more. oh well I guess it means i can save my money for a gpu.

 

[InnocentCriminal]

People are forgetting this is a really early look at a Nehalem based set up. Give it a month and if they repeat this but with a updated motherboard [BIOS] I expect it'd be a completely different story. No, it will be a completely different story.

 

[HTC]

I would suggest a modification to the OP title: it's favoring the 4850 when it should be favoring the Nelahem CPU scores.

Something like: "A preview of Nelahem CPU scores, using a HD4850".

 

[btarunr]

I would suggest a modification to the OP title: it's favoring the 4850 when it should be favoring the Nelahem CPU scores.

Something like: "A preview of Nelahem CPU scores, using a HD4850".

The video-card doesn't affect CPU score. The only reason the card is mentioned there is to show the scores are pure CPU scores and no possibility of NVIDIA adding PhysX score to the CPU score. I'm not the only one using this kind of title, even the likes of Nordic Hardware are, Google out to find more.

 

[HTC]

The video-card doesn't affect CPU score. The only reason the card is mentioned there is to show the scores are pure CPU scores and no possibility of NVIDIA adding PhysX score to the CPU score. I'm not the only one using this kind of title, even the likes of Nordic Hardware are, Google out to find more.

Yes, i know.

It's still hyping the VGA when it should be hyping the Nehalem CPU scores (i'm referring to the title: nothing else), wouldn't you say?

I'm aware that physX adds a LOT to the CPU score which is why they used an ATI card.

Personally, if any physX thing were to be used (Ageia card or those modified CUDA drivers (?) that showed a 3850 with a CPU score of 22K+ in a Vantage P score), the CPU score would be WAY higher then 17996.

How about "Nehalem benchmarks: CPU scores using a HD4850"?

 

[btarunr]

Yes, i know.

It's still hyping the VGA when it should be hyping the Nehalem CPU scores (i'm referring to the title: nothing else), wouldn't you say?

I'm aware that physX adds a LOT to the CPU score which is why they used an ATI card.

Personally, if any physX thing were to be used (Ageia card or those modified CUDA drivers (?) that showed a 3850 with a CPU score of 22K+ in a Vantage P score), the CPU score would be WAY higher then 17996.

How about "Nehalem benchmarks: CPU scores using a HD4850"?

Alright, the user always comes first. Changed title, vBulletin limitation shows old title, new one on the main page. Your feedback is always welcome.

 

[HTC]

Alright, the user always comes first. Changed title, vBulletin limitation shows old title, new one on the main page. Your feedback is always welcome.

You didn't have to remove the "HD4850" from the title but, personally, i think it's better now as opposed to before.

Plus, in theory, it should avoid more posts saying something like "the graph card should have been another one because this ... or that ..."


IMHO, when you trying to enphasize something (in the body of the post) and hype (wrong word: can't think of a better one, right now) another on the title, it may get confusing and something similar to the "Mattgal thing" is bound to occur.

 

[btarunr]

You didn't have to remove the "HD4850" from the title but, personally, i think it's better now as opposed to before.

Plus, in theory, it should avoid more posts saying something like "the graph card should have been another one because this ... or that ..."


IMHO, when you trying to enphasize something (in the body of the post) and hype (wrong word: can't think of a better one, right now) another on the title, it may get confusing and something similar to the "Mattgal thing" is bound to occur.

Mattgal is using a flawed logic that the 'GPU(s) will 'bottleneck' CPU score and that they should've used some el-fatto graphics hardware and obtained better scores.

 

[HTC]

Mattgal is using a flawed logic that the 'GPU(s) will 'bottleneck' CPU score and that they should've used some el-fatto graphics hardware and obtained better scores.

Yes, but i was referring to him/her as an example only.

Please re-read posts #2 and #3 of this thread: a classical example of confusion between what was in the title and what is in the body of the OP.


Too bad vBulletin limitations still show the old title, instead of the new one but, as far as i know, that can't be helped

 

[Morgoth]

hmm wierd i posted this in ati grapics section and non post
and here its gets flooded and complaining abouth the score :S
bloomfield eats your Penry clock for clock!
btw for that nvidida fan boy thats talking abouth the bottleneck
There is still no infomation if Nvidia is allowed to make chipsets for bloomfield

 

[HTC]

hmm wierd i posted this in ati grapics section and non post
and here its gets flooded and complaining abouth the score :S
bloomfield eats your Penry clock for clock!
btw for that nvidida fan boy thats talking abouth the bottleneck
There is still no infomation if Nvidia is allowed to make chipsets for bloomfield

You mean this post: i had seen it, yes.

It's all because of the title. Right now, HD48x0 are "on fire", so to speak, and they attract many viewers. Bloomfield is still somewhat unknown to many, myself included.

 

[Steevo]

HT only works well if you have the on die cache to feed the actual core, or enough memory bandwidth to supply data, but if you have that then the program in wait state is either poorly coded or the amount of time spent switching tasks and getting the data for the next task in line will be greater than the amount of time spent getting the current tasks data.


HT was a waste of resources in the p4 days and it still is.

 

[Morgoth]

no its still not ht works great on my p4