Another review for the list on the 1st page:
http://www.phoronix.com/scan.php?pag...ynnfield&num=1
This one is a review under Linux and results are quite surprising(no amd quad core used,but instead a X3 Deneb)
Printable View
Another review for the list on the 1st page:
http://www.phoronix.com/scan.php?pag...ynnfield&num=1
This one is a review under Linux and results are quite surprising(no amd quad core used,but instead a X3 Deneb)
I would not call this thing "review". Theirs performance sum based only on 4 tests - winrar, 3dmark, vantage and superpi (and seems that winrar have the most weight in the sum). I suppose we should keep the thread clean from such "reviews"Quote:
Originally Posted by informal
.
Yeah, I'd be satisfied with just seeing 58xx review, should be very soon I guess!Quote:
Originally Posted by informal
I posted another one with a 26 applications showing almost the same numbers.Quote:
Originally Posted by kl0012
BTW ,under linux,even x3 710(2.6Ghz) is holding its own while going against the new i5 and i7-look at phoronix review under ubuntu(that's real world workloads). That's 3 cores @ 2.6GHz against 4c/8t @ 2.66-3.46Ghz :)
I have to agree, "reviews" that gives the perception that PIIs are better than what it is... and for some reason these "reviews" are only put up by a few users who have anti-Intel agendas. :rolleyes:Quote:
Originally Posted by kl0012
Ok as it stands i think the performance in high bandwidth CF "Like 4870 x2 CF" will be:-
i7+x58>PhII+790>i5+P55
For someone looking to build a screaming mid-range machine, it's hard to argue with i5 and P55. Tons of performance, value etc.
Also, the motherboard selection is fantastic!
Can't agree. The second diagram (which excludes games tested on high res with 9800GTX) actually shows that i5-750 overal faster then Pii-965.Quote:
Originally Posted by informal
Hm... I wonder why they used Ubuntu 9.10 Alpha release. Alpha releases usually includes many debug code and has unpredictable performance on different systems.Quote:
BTW ,under linux,even x3 710(2.6Ghz) is holding its own while going against the new i5 and i7-look at phoronix review under ubuntu(that's real world workloads). That's 3 cores @ 2.6GHz against 4c/8t @ 2.66-3.46Ghz :)
Well we can go all day long finding errors in testing or test choices. 750 due to its high turbo mode looks faster than 965BE,but without it it's a bit slower.Also with turbo on,750 is a bit faster or on par with 860 i7 too... The difference between i5/965 and top of the line 870 with turbo and smt on is 16 percent in office and multimedia apps... No gaming involved. 16% is decent but nothing to write home about,especially if you consider the difference in price.Quote:
Originally Posted by kl0012
Phoronix website is quite professional all around. There is a reason they used that release.They tested the 920 i7 quite a long while ago and were quite pleased with its performance. You can always ask them though :).
pfff Kyle tested Lynnfield under gaming @ 640x480 what a champ!
:down:
http://hardocp.com/article/2009/09/0...7_processors/5
Yea, what a champ... this is a CPU review not a GPU review. :rolleyes:Quote:
Originally Posted by Nedjo
So you suggest that car reviewers pull out engine out of two ton car and put it on bench table to give you acceleration performance torque ratio, fuel efficiency etc... :rolleyes:Quote:
Originally Posted by Clairvoyant129
that page is about "Lynnfield under gaming", and people in 2009 don't game on SMARTPHONE resolution!!
Kyle, and obviosly you don't have clue about CPU role in gaming... if you did have clue you would know that true stress for CPU under gaming is with more detail!
The point is to find sweet spot that would stress CPU, and still not represent GPU bottleneck!
Ideal CPU under gaming conditions would be testing with Dual GPU (1 gig cards) setup (driver overhead comes as a burden for CPU) @ 1680x1050 Medium/High settings!
So u want to get the i5/i7 and play at 640x480??? Ohh okk so the logic is you get more frames...ohh :up:Quote:
Originally Posted by Clairvoyant129
EDIT: Where the hell are all the mobo reviews?? Wanted a review of P7P55D Plain and P7P55D LE soo bad.....
Well in real function that would make the cpu faster, once again turboboost is an integral function of the cpu now as much as speedstep to lower the speed. If people are looking at the review to decide which out of the box cpu is faster then it is what it is, what they see is what they get.Quote:
Originally Posted by informal
Do you expect them to clock the i5 at the same speed as a comparably priced AMD cpu to be fair...
I don't think it is even worth to discus that. I prefer to wait for official release of Ubuntu 9.10 (somewhere in October) with all bugs filtered out (especialy those affected new hardware). I see no point in testing cpus on different alpha version of unstable software.Quote:
Originally Posted by informal
Is this what you're looking for?Quote:
Originally Posted by Nedjo
http://xbitlabs.com/articles/cpu/dis...0_9.html#sect0
Are you serious? As mentioned, he's not stressing the GPU. The lower the resolution, the more CPU bound the game is.Quote:
Originally Posted by Nedjo
If it were up to you, you would probably want to test it with AA also.:rolleyes:
Gaming with more detail is more stressful for the GPU, not so much the CPU. In case you didn't understand, he's trying to isolate the CPU performance.
Jeez... did you even read the text of the article?Quote:
Originally Posted by Nedjo
I quote Kyle (bold added by me):
You may not find the information useful, others might. Personally, I don't care as I don't game, but putting down a testing procedure without at least some basic research is wrong in my opinion. Not only does Kyle state his goal very clearly and percisely, he also raises the necessary caveats on how the data can be used.. Again, if you don't find the data useful move on. However, I don't know of any other way to get the answer to the question he is raising. Which is, with as little GPU bottleneck as possible, just how fast can the CPU go on a particular game?Quote:
As always, these benchmarks in no way represent real-world gameplay. They are all run at very low resolutions to try our best to remove the video card as a bottleneck. I will not hesitate to say that anyone spouting these types of framerate measurements as a true measuring tool in today’s climate is not servicing your needs or telling you the real truth.
The gaming tests below have been put together to focus on the processor power exhibited by each system. All the tests below consist of custom time demos built with stressing the CPU in mind.
Its only game when HT will gain some speed (multicore game..). Like someone said its good way to check CPU power with low res...Quote:
Originally Posted by Nedjo
hmm so i5 has made AMD's mid range processors incompetent once again. I wonder what AMD will do to counter this? I hope whatever they do is better for themselves and consumers as well.
interesting. nehalem is 10% faster clock for clock in memory intensive apps.Quote:
Originally Posted by highoctane
Why don't you explain how prefetchers work and when they work? :rolleyes:Quote:
Originally Posted by Clairvoyant129
While doing that you can also explain the differences in how the cache work on i5/i7 vs Phenom II
It is one thing to just say someone don't know, but to explain is another ballgame
Once again: as the scene that's being rendered simpler - there's less work to be done for CPU - less geometry to feed GPU, less traficking over the PEG, less driver overhead to deal with...Quote:
Originally Posted by Salavat23
I agree that there's a point in testing CPU's where GPU is becoming bottleneck and it shouldn't be passed for the sake of shadowing CPU performance, but going so low as Kyle did is ewen worse 'cos it misleads readers about true potential and usability of CPUs in gaming scenarios
ABSOLUTELY! If I'm spending serious money on new powerful HW I want to see how it performs on my big 30" TFT, and apparently here's the answer:Quote:
If it were up to you, you would probably want to test it with AA also.:rolleyes:
http://www.anandtech.com/cpuchipsets...spx?i=3634&p=9
2560x1600:
http://www.abload.de/img/anandtechintelscorei78iueo.jpg
He don't have clue... or maybe does... anyhow he's misleading readers! If he wanted to stress out CPU in gaming he should have run software rendering of Unreal 1 and Quake 1 in 640x480!! :rolleyes:Quote:
Gaming with more detail is more stressful for the GPU, not so much the CPU. In case you didn't understand, he's trying to isolate the CPU performance.
Insulating CPU in trying to portrait gaming performance is by definition nonsense, but you can make of it something if you know how to do it, and I've suggested how! And as someone pointed xbit is using that approach, as well as anandetch... http://www.anandtech.com/cpuchipsets...spx?i=3634&p=8
but yeah Kyle is smartest of them all :rolleyes:
How is anand misleading? At that res you get very GPU limited very quickly. Ok it's Crysis but you know the further down the ladder of graphical fidelity you go the more cpu intensive it is, hello Source engine, hello COD4/5! It's a strange world in that if you have a 30" you are better off spending your cash on graphic card upgrades rather than cpu upgrades. The smaller your monitor the more cpu makes a difference, the bigger it is the less it does. So at the highest res available to consumers, you're fine with a high clocking e8x00. The sweet spot is 1920x1200, the cross section where you get most bang for buck from both. I've been considering downgrading for some time, but hopefully the 5xxx series and GT300 will change the noodle. :)Quote:
Originally Posted by Nedjo
who said anand is misleading? read again my post... anand's and xbit's approach is much more informative than {H}ocp'sQuote:
Originally Posted by tdream
the smaller the monitor - smaller is resolution so my advice to people with smaller monitors is go with AA as high as possible!
I'd agree with you that this and next year will establish 1080p as a mainstream resolution that's covered with a range from 22 'till 26 inch monitors, and that's a resolution that should be targeted in reviewing!
When it comes to methodology, my POV is to tune down any Shader, or other special effects, to leave any physics-type of effects as highest as possible...
point is to burden CPU as much as possible with all the tasks it should be doing to the max, and on the other side not to bottleneck graphics card... going on VGA and SVGA resolutions isn't answer... as simple as that!
Answer is MultiGPU system with tuned low Shader effects, and upped high geometry complexity ( 'cos CPU feeds geometry) and texture quality (modern engines are streaming engines so PCI Express traffic, and implementation counts) in 1080 rez!
All my respect onethreehill :up: