• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Test upon Intel Core i7

Just so you know, neither your Sempron nor Core i7 has a FSB. They have high speed direct interconnects (HTT and QuickPath respectively).
 
bestelement said:
I believe Intel Core i7 is a very brilliant CPU. The reason why we don't usually see it that way is we have been so familiar with Intel CPU's great power and overclockability. Compare Core i7 with any other CPU at present, it is often the better, but not in games, at lease now.

Core i7 is in its somewhat beta or RC course of the development, so I think further improvements will be done before they get launched.

And mentioning overclocking ability, that site had posted some info.

http://uneit.com/2008/09/24/a-little-test-of-intel-core-i7s-overclockability/


Also, a question bewildering me. Why on earth that everytime we are nearing a big release, infos are always leaked from those Oriental countries. Isn't it the fact Americans designed those?
Click to expand...
Bah, they blanked out the multi and freq so we can't tell what they toyed with. It's an ES so it likely has an unlocked multi, so that might not be a good indication of the overclockability we will see.
 
bestelement said:
Also, a question bewildering me. Why on earth that everytime we are nearing a big release, infos are always leaked from those Oriental countries. Isn't it the fact Americans designed those?
Click to expand...
the mobo manufacturers are all oriental.

WC Annihilus said:
Bah, they blanked out the multi and freq so we can't tell what they toyed with. It's an ES so it likely has an unlocked multi, so that might not be a good indication of the overclockability we will see.
Click to expand...
Its a 920, 133x20 base clock. ES, so probably WAS unlocked.
 
Weird, when I first read this I could have sworn that it was a 940 with a 22x multi and "fsb" freq of 133mhz. Unless I have my i7 articles mixed up.
 
samsonJS said:
Weird, when I first read this I could have sworn that it was a 940 with a 22x multi and "fsb" freq of 133mhz. Unless I have my i7 articles mixed up.
Click to expand...
The benchmarking tests are a i7 940 (2.93, 133x22), the overclocking test was a i7 920 (133x20).

At least we have a solid naming scheme already. The lesser cores will be i7 less than 900 series. 10 points per multi (133Mhz). 5 points for extreme edition. An i7 chip at 3.46 would be "i7 980". Maybe a 990, @ 3.6Ghz. Which may be the highest they expect the generation to go... :( No real room to increase the base clock either.

Im curious as to why [intel] picked 133 instead of 200...
 
You can't look at the same way anymore. And this was Intel can sell new products as time goes by.
 
Overall, the i7 for new people who have to have the latest and greatest will be let down as things won't be optimized for the i7 architecture.

Going from my P4 3.2E to my Q6600 there was a night and day difference.

People going from a C2D or a C2Q to the i7 chipset aren't going to be nearly as impressed.

Once DDR3 memory starts to become mainstream next year, prices will drop and the adoption rate will increase slightly. Game wise though, I think things are going to be about the same for the next 18-24 months.
 
Digital Viper-X- said:
yes but what new instructions is I7 bringing that penryn didnt have? that would affect gaming that is
Click to expand...

It's not just about what instructions it supports, but also how they are implemented.
Just the fact that i7 has a different cache hierarchy can already result in code optimized for Core2 caches performing sub-optimally. Another thing is that the Core i7 has massively more memory bandwidth than the Core2. This means that algorithms that were previously optimized to minimize bandwidth, might now be faster when you use more bandwidth, but less instructions (skip code that tries to compact the dataset).
x86 is not just x86, you know.
One example is the Pentium Pro, which handled partial register operations poorly. This meant that a lot of hand-optimized 486/Pentium code could perform really poorly, and had to be rewritten to avoid partial register operations.
Another example is the Pentium 4, which had a completely different architecture from the Pentium 3, and required code to be optimized in a completely different way.

Not that I would expect a difference as dramatic as these examples, but still I think there is a difference between optimal code for Core2 and optimal code for Core i7.
 
Atech said:
I would like to se min/avg/max graphs on final hardware and some none-GPU limted (some games actually DO stress the CPU pretty hard)examples, before making such a statement.
Click to expand...

you and me both, theres not alot of info out there. my statement was based upon what was in the article.
 
Scali2 said:
It's not just about what instructions it supports, but also how they are implemented.
Just the fact that i7 has a different cache hierarchy can already result in code optimized for Core2 caches performing sub-optimally. Another thing is that the Core i7 has massively more memory bandwidth than the Core2. This means that algorithms that were previously optimized to minimize bandwidth, might now be faster when you use more bandwidth, but less instructions (skip code that tries to compact the dataset).
x86 is not just x86, you know.
One example is the Pentium Pro, which handled partial register operations poorly. This meant that a lot of hand-optimized 486/Pentium code could perform really poorly, and had to be rewritten to avoid partial register operations.
Another example is the Pentium 4, which had a completely different architecture from the Pentium 3, and required code to be optimized in a completely different way.

Not that I would expect a difference as dramatic as these examples, but still I think there is a difference between optimal code for Core2 and optimal code for Core i7.
Click to expand...

the only thing I can really see affecting performance in this case, is the cache hierarchy. but I don't understand why it would boost performance in everything other then games? are games more cache intensive then the other apps tested?
 
Digital Viper-X- said:
the only thing I can really see affecting performance in this case, is the cache hierarchy. but I don't understand why it would boost performance in everything other then games? are games more cache intensive then the other apps tested?
Click to expand...

You can't generalize for all games. An RPG has massively different requirements from a arcade racing game, for example.
Therefore I think it's more likely that the bios/chipset aren't quite up to snuff yet, and suboptimal communication between CPU and GPU is what's holding these games back a bit. The other applications aren't graphics-intensive, so you get to see the 'raw' CPU performance, which seems to be equal to or better than a Core2 in every case.
Therefore it's likely that games will perform equal or better aswell, assuming the rest of the system also performs equal or better than a Core2.

Now there is a possibility ofcourse that the integrated memory controller makes certain operations less efficient than in the old Core2 configuration. We'll have to wait and see.
 
S-B said:
I want to see FSX benched.
Click to expand...

Why, MS cant eve get a decent working service pack out to properly enable the use of multiple cores on current systems, Nah. wont make it run any better - read the articles.

Intel is saying screwing gaming, lets make a kick ass chip for an industry we probably make 50x fold on profits then gaming - SERVER! market! and workstations!
 
I think it's a complete misconception anyway that you would actually design a processor for a certain purpose. It's not called a general purpose processor for nothing.
You can't design a CPU that is 'good at gaming'. It doesn't work that way.
Game engines are incredibly complex and consist of many different parts, basically requiring a CPU that is 'good at everything'.
Athlon64 was a good gaming CPU because it was 'good at everything'. Core2 is a good gaming CPU because it is 'good at everything'.

Now Nehalem may have some features that aren't useful outside the server/workstation market, but it's complete backward logic to say that they put those in to 'screw gaming'.
Athlon64 has some of the same features, but I never heard anyone complain that they were poor gaming CPUs because of it.
These features won't actually hurt gaming. They just won't benefit gaming. At least, not current games. It's entirely possible that future games take advantage of the extra memory bandwidth and hyperthreading, and then suddenly Core2 will look like a lesser gaming CPU, just as Athlon64 is now.
 
SymptomsReagent said:
Overall, the i7 for new people who have to have the latest and greatest will be let down as things won't be optimized for the i7 architecture.

Going from my P4 3.2E to my Q6600 there was a night and day difference.

People going from a C2D or a C2Q to the i7 chipset aren't going to be nearly as impressed.

Once DDR3 memory starts to become mainstream next year, prices will drop and the adoption rate will increase slightly. Game wise though, I think things are going to be about the same for the next 18-24 months.
Click to expand...

im on a p4. 3.2gh. there is no point in me getting a q6600. im going to get nehalem to help my pc last 3 years with just upgrading maybe ram and gpu once in a while to keep up with gaming and general APP usage
 
Nehalem for game/general app? That sounds like overkill since you don't seem like you upgrade much, unless you're talking about Nehalem after launch and when prices settle down.
 
Guess so, but what "general apps" and games will get a boost from i7? Think we all know now that i7 won't do much for games, and general apps.. well, single cores can do that and 2/4 cores are more than enough. So imo a basic system doesn't need overkill when the overkill does not provide any benefit. Then again it's not even out so who knows what kind of rabbit will come out of the hat; all we can use are preliminary results.
 
Future-proofing, I suppose.
We all know that at some point, we WILL require that faster processor and that extra memory etc.. Else we'd still be using 8086 systems today.
I always do it... When I upgrade, I tend to buy the fastest system I can afford (without going overboard on Extreme Editions or other poor value-for-money niche products), then trying to ride it out as long as I can. At least it means I can run all my applications and games very comfortably until eventually the bottom drops out.
If I try to go for the most 'cost-effective' system when I'm upgrading, I know that the bottom can drop out any minute. Getting a fast system will usually keep me in the upper regions for the first 1-2 years. Heck, my E6600@3GHz is still chugging along nicely. I can probably skip Nehalem, and wait for the next shrink.
 
Yea I'm with you, skip Nehalem and wait for its successor that will do what I want it to do, like load WC3 maps faster :D
 
Scali2 said:
I think that's the point... Get an 'overkill' system now so you don't have to upgrade for years.
Click to expand...

exactly. i dont upgrade often as u see so a nehalem will last me 3+ years by just chopping and changing gpu's and maybe memory if games need more then 2gb memory which im going to start off with :)

yes i know 4gb is the norm but im fully aware of the costs of ddr3 ram so 2gb i7 will do me just fine till ddr3 becomes dirt cheap and i can go from 2gb to 6gb triple channel baby YEAH:D
 
I upgrade constantly. I'll probably get the Core i7 but of course the real decision won't come until they've been released and I can see the results of the performance tests for myself. Based on existing benchmarks I probably wouldn't buy it today because of the sorely lacking gaming performance.
 
Dan_D said:
I upgrade constantly. I'll probably get the Core i7 but of course the real decision won't come until they've been released and I can see the results of the performance tests for myself. Based on existing benchmarks I probably wouldn't buy it today because of the sorely lacking gaming performance.
Click to expand...

cool i understand. i probably wouldnt if i upgrade yearly too. imo i7 is good for people who dont upgrade much.

want a pc that will last u a while, get a i7
 
I'll be rooting for Westmere or Sandybridge V=.=. Wasn't it said before that Nehalem might all become socket 1166 or w/e as it moves away from release? That'd mean that 1366 won't be compatible w/ 1166, so early adopters would need a new mobo? Haven't seen much discussion regarding this, all are just speculation.

intel release schedule btw: http://www.techpowerup.com/68544/Nehalem_s_Successors_Caught_on_Slides_at_IDF.html
 
Wow, I would have hoped it could have done better in games. If you average the results in the review its close to a QX9770 at the same speed.:confused:
 
Is it out of the realm of feasibility to assume improved chipset drivers could improve performance; specifically related to games?
 
the best thing about these new chips is the single socket. No more lga775 and 771;I can get a nice dual socket mobo and upgrade till I am blue in the face. I am just gushing thinking about my 12core vm rig.
 
TwinTurboJosh said:
Is it out of the realm of feasibility to assume improved chipset drivers could improve performance; specifically related to games?
Click to expand...

No, it's quite possible.
Then again, there is a chance that the current results are as good as it gets. We'll just have to wait and see.
 
Scali2 said:
No, it's quite possible.
Then again, there is a chance that the current results are as good as it gets. We'll just have to wait and see.
Click to expand...

Exactly. Could get better, it might not.
 
You must log in or register to reply here.
Back
Top