4GHz is still slower than i7 870 @ 3.5GHz in Vantage and openGL CinebenchR10 with a 5970
Printable View
4GHz is still slower than i7 870 @ 3.5GHz in Vantage and openGL CinebenchR10 with a 5970
I don't understand what you're saying here.Quote:
Originally Posted by initialised
An 870 @ 3.5 rated with turbo would be running a 160 clock and going 3.8 on all 4 or 4.1Ghz in single or dual.
An 870 @ 3.5 with no turbo would be running the same 160 clock and not changing.
A 1055T running at 4.0 would be running about 286 clock and doing 4.7 single, dual or triple. (I don't think we've seen this yet.)
A 1055T running at 3.4 would be running about 242 clock and doing 4.0 single, dual or triple.
Which combination of those is being compared here?
wtf clock for clock thuban will beat i7 quadsQuote:
Originally Posted by initialised
Winrar is a little fail for X6
Vs i5 750/860/870 stock
http://lab501.ro/wp-content/uploads/2009/09/WinRar.jpg
Vs i5 750/860 at 4ghz. Phenom X4 965 is at 3.8ghz.
http://lab501.ro/wp-content/uploads/.../WinRar-OC.jpg
http://lab501.ro/wp-content/uploads/...ma-580x215.png
http://lab501.ro/procesoare-chipsetu...d-in-actiune/8
1055T prices already appeared in Romania about 15-20$ more than an i5 750 to be in stocks in 4 days
It's clearly for me that in some application 1055T would be a little better than i5 750, in some closer to i7 930/860, may be equal in a very few. Games are a big question, smaller cache per core might be a problem.
All in all 1055T a very good processor price/performance ratio a think that 1055T is the real concurence for both i5 750 and i7 860/930, 1090T is vey expensive about 300$ in Romania, 25-30$ more than an i7 930/860. It's 100$ more exepensive than 1055T...
diferent is pracice using and benchmark, Phenoms like unzip small files (winrar or winzip decompression with much small files)
Also winrar bench just get excited with "threads". In real life use the numbers from benchmarks (winrar/7-zip) are quite different. The same goes to 3dmark vantage.
Now with programs that take full advantage of cpu power(rendering, video compression) i7 quad and x6 fight neck to neck.
lol xdan you are actually comparing a 750 with 1055?? Are you even half serious? First of all in real MT workloads(winrar built-in test is just a BS benchmark) Thuban will not only kill i5 but will probably edge out i7 as well.In single thread workloads 750 was actually faster than Deneb thanks to its very aggressive Turbo mode apart from slight IPC advantage without it.Now Thuban levels the playing field by having similar tech to its disposal. 750 was already competing with Deneb and will continue to do so,while Thuban is an i7 territory.Latest games have always been shader bound and cache sensitive and Deneb already did great in high resol.+high detail settings(you know the ones people who buy high end HW actually use in real life:rolleyes:,not 1024x768 with NoAA and NoAF).
I don't think those benchmarks are valid. They aren't at all consistent and don't jive with what can be observed in real-world WinRAR performance. The 965 gaining 4.6% for a 23% clock increase is just silly.Quote:
Originally Posted by xdan
If those numbers were actually taken from WinRAR's built-in benchmark tool then the tool is pretty worthless as it isn't indicative of performance that anyone will ever experience.
I listen of a new model better than the 1090T, the 1095T...anyone know that is true?
http://lab501.ro/wp-content/uploads/...ma-580x215.pngQuote:
I don't think those benchmarks are valid. They aren't at all consistent and don't jive with what can be observed in real-world WinRAR performance. The 965 gaining 4.6% for a 23% clock increase is just silly.
It's not a 23% clock from 3.4 ghz to 3.8ghz, 400Mhz are about 12%
An my mistake i5 750 i7 860/975 are at 4ghz.
And the X6 at about same clock with higher NB and memory scores 800-900( about 30% more) points more with 50% more cores.
Will see but think that L3 cache per core is 25% less, and the IMC will bottleneck as usually probably even more 6 core than 4 core. Hopefully Buldozer will have a 2X better IMC.Quote:
lol xdan you are actually comparing a 750 with 1055?? Are you even half serious? First of all in real MT workloads(winrar built-in test is just a BS benchmark) Thuban will not only kill i5 but will probably edge out i7 as well
My estimations are that 1055T will be at middle between i5 750 and i7 860/930 i mean especially in application that make the difference between AMD and Intel like Photoshop, Winrar, video encoding, etc.
There will be some application in wich 1055T will be equal to a 930 or slight better but few.
I'm not talking about the 1055T benchmark you posted. I'm referring to the the 965 versus i7/i5 comparisons that I actually quoted. Those images show all the chips compared at stock and them again at 4.2Ghz. 3.4 to 4.2 is a 23.5% increase.Quote:
Originally Posted by xdan
Furthermore, we've clearly established that WinRAR's benchmark tool is worthless.
X4 965 is at 3.8ghz( C2 rev) see in the picture above.
i5 750, i7 860/975 are at 4ghz.
My mistake again.
http://lab501.ro/procesoare-chipsetu...d-in-actiune/8
Review made by team lab501 , Romania.
Ok, so you're showing a table that has the Intel Processors, clocked @ 4GHz w/ 2000MHz DDR3, while the AMD 965 is @ 3.8 w/ 1600MHz DDR3. Anything fishy about that?Quote:
Originally Posted by xdan
We can already see from ridney's results that a 1055 @ 3.8Ghz is about 4% faster in x264 than an i5 750 at the same clock and the other Nehalem's don't gain a whole lot in that benchmark from HT.Quote:
Originally Posted by xdan
Photoshop does seem to be fairly memory dependent and doesn't appear to scale very well with additional cores so it wouldn't be surprised if the X6 doesn't do very well at all. My own calculations put the theoretical performance of a stock 1055T at somewhere around that of a stock 955.
Even with rev C3 X4 955/965 can do max 4- 4.1GHZ and about 1800 memory for 24/7 and so i don't see your point.Quote:
Intel Processors, clocked @ 4GHz w/ 2000MHz DDR3, while the AMD 965 is @ 3.8 w/ 1600MHz DDR3. Anything fishy about that?
I can show this too...
http://i967.photobucket.com/albums/a...24m_XS_Air.jpg
Vs
http://forum.lab501.ro/attachment.ph...67&d=970068612
860 is 100Mhz higher let's say that that clock per clock are equal.
So what youre saying is, 1055T which costs the same as i5 750 looks to be on par with i7 860 ,has lower platform cost and longer platform longevity(am3 is BD compatible) .
Looks like we agree ;-).
And theres of course 1035T, which is going to cost even less (est 179$).
...
And there is 960T ;-) ,which can be unlocked to six core ;-).
another good one.. u might have to run w/ ur i5 to compare
http://hostimage.co.uk/08-Apr-10_0F69-4BBDB0E5.jpg
http://i70.photobucket.com/albums/i82/mrjohnson2/vv.jpg
My i5 750 Wprime at 4.4ghz
http://forum.lab501.ro/attachment.ph...0&d=1268819122
Umm.. have you decided to run any stability testing overnight or something ? I want some 24/7 stable clocks posted.
Why you even compare your i5 @ 4.4Ghz with Thuban @ 4.3Ghz when they are clearly no match(Thuban crushes it silly...). Thuban is even faster than i7 in wprime clock for clock.
Somebody ask me to run i5 to compare, i know that isn't competition in Wprime for X6...
For general information my i5 750 was stable at 4.2ghz - 1.39v. Now i have an i3 530...
To prime more than 2 hours...overnight prime sesions is just nonsens.
Quote:
Originally Posted by Praz
Quote:
Originally Posted by initialised
click for larger image
http://www.edgeofstability.com/image...4000_lin_s.jpg
Praz ,nice stability there man :) . Thuban is really something special.
Praz, im seeing there you got x15 multiplier working on 1035T (base multi of 13x).How did you do that ? :-)
Turbo glitch maybe?
If you look at the screenshot it says 6 threads on linX, and finished, so he probably has turbo disabled.
And he was proving stability at 4ghz, which he could not do with 13x and 266 HTT ...
My guess was that maybe he can set 15x through software.
or maybe that's not a 1035T but a 1075T which runs at 15.0x 200? anyways, great stability result.
Still waiting on word from the higher ups. Not sure if it the chip or the board/BIOS that is allowing the Turbo multipliers to be set.Quote:
Originally Posted by RaV[666]
No glitch.Quote:
Originally Posted by informal
Turbo was disabled. The multiplier is set to x15 in the BIOS.Quote:
Originally Posted by RaV[666]
No 1075. It is a 1035.Quote:
Originally Posted by ridney
that's good news, you can actually use the extra turbo multi when it's disabled. nice :)
Thats what i was saying from the get-go! Probably in some bioses it will be blocked as by AMD guidelins but its very much possible it will be software settable!
1035T here i come :P
ha! support AMD and get the 1090T :ROTF:Quote:
Originally Posted by RaV[666]
I always get their slowest one from highest bin :P ,it has become my tradition of being a cheapskate :PQuote:
ha! support AMD and get the 1090T
Guys don't get too excited yet. I'm not sure this is what one can normally expect which is why I asked for clarification.Quote:
Originally Posted by RaV[666]
There is only one sure thing that would prevent it, is your chip an ES ? Because if its not, if its the same as retail, it will be possible.Quote:
Guys don't get too excited yet. I'm not sure this is what one can normally expect which is why I asked for clarification.
Can you check with k10stat and amd overdrive which multis are available to you with both turbo on and off ?
LinX cool! Thanks, what about max clock in LInx with 1.5V? Maybe 4100Mhz?
The only utility that have look at is HWINFO. It shows default multiplier to be x13 and max x15.5. x15.5 is as high as I can manually set.Quote:
Originally Posted by RaV[666]
4100MHz may take less voltage then 1.50. I have the CPU in the Crosshair IV now and it's running 4060MHz with VCORE set to 1.42V. That's with 300MHz bus frequency and both NB and HT frequency at 3000MHz.Quote:
Originally Posted by FlanK3r
Praz:can i have question? Im tired (a bit) waiting at new Gigabytes motherboard, but have fear from ASUS. Do u have not problems with BIOS voltage and real voltage in CPU-Z at windows? How much diferent? Gigabytes are great, i set 1.425V, in windows i have 1.425, no drop or overvolting etc. Thx.