R520 Benchies at HA

[ddogg]

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

well it looks like you didnt read my post correctly...i didnt say it was possible i just said what-if like everyone here says what-if it were 24pipes or 32pipes blah blah!!!!

 

[Polish3d]

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

Actually with a voltmod and dual slot, I doubt it would be that difficult.

My GTX reports as 515 max, 75mhz higher isn't much, and I'm only 40mhz away from 1400, stock volts, stock cooling. IIRC a guy with 615/1450 got 11k in '05

 

[nitromullet]

Originally posted by: munky
It would be funny if this whole uproar caused Ati to paper-launch the card early (like tomorrow ), and we'd have some official benches (hopefully less disappointing than these).

I don't think that will happen, nor do I think that the real numbers will be significantly different that these. Obvioulsy, that's just my opinion. Honestly, ATI's reaction made these benchmarks more credible. Had they just said they were fake, it might have made sense, but when they said they had not shipped any X1800XT cores to their board partners three weeks before their launch... There are only two legit reasons why they wouldn't have shipped X1800XT cores:
1) They don't have enough
2) ATI is going to make the first batch of cards themselves. This is certainly a possibilty, but I don't think that would be a good indicator of high availability.

 

[Munky]

Originally posted by: Frackal

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

Actually with a voltmod and dual slot, I doubt it would be that difficult.

My GTX reports as 515 max, 75mhz higher isn't much, and I'm only 40mhz away from 1400, stock volts, stock cooling. IIRC a guy with 615/1450 got 11k in '05

You have to look at the percentage increase. 75 mhz on a 3 ghz cpu is not much. 75 mhz on a gpu that's already pushing the limit at 515 is a pretty big jump, actually.

edit: and that's 85 actually, 515 + 85 = 600

do you have a link to the guy who got his to 615?

 

[Keysplayr]

Does anybody know if the X1800's are fabbed with low-k?

 

[ddogg]

Originally posted by: keysplayr2003
Does anybody know if the X1800's are fabbed with low-k?

ya i think i saw the inquirer post that it would be fabbed with low-k although im not 2 sure.

 

[BFG10K]

Based on those benchmarks my next card will be a 7800 GTX. Of course a minor miracle could happen and the benchmarks could show something different in a month's time.

 

[ArchAngel777]

Originally posted by: thatsright
As soon as I saw the words "pre-release drivers," I just stoped reading.

Thats 5 minutes of my life I'll never get back

Yeah, because released drivers predate pre-released drivers... Uhhh... Think about that for a minute. Pre-Release does not mean the drivers are inferior, it simply means they have not been released. It can mean they are inferior, but it doesn't automatically mean that. Why are they considered pre-release? Because the cards have not shipped yet.

Anyway, if that was enough to want your 5 minutes back, well... Uhm, I do not know what to say.

 

[Elfear]

Originally posted by: munky

Originally posted by: Frackal

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

Actually with a voltmod and dual slot, I doubt it would be that difficult.

My GTX reports as 515 max, 75mhz higher isn't much, and I'm only 40mhz away from 1400, stock volts, stock cooling. IIRC a guy with 615/1450 got 11k in '05

You have to look at the percentage increase. 75 mhz on a 3 ghz cpu is not much. 75 mhz on a gpu that's already pushing the limit at 515 is a pretty big jump, actually.

edit: and that's 85 actually, 515 + 85 = 600

do you have a link to the guy who got his to 615?

Here you go munky. Lots of guys on the orb are running some crazy numbers. (Hit "Show me the projects", I guess it doesn't populate the results from the link)

ViperJohn, the volt-modding guru, told me that a his air-cooled cards get anywhere between 540-600Mhz on the core and one of his watercooled cards ran at 640Mhz artifact free.

I wish one of the guys with one of these uber cards would run some gaming benchmarks. I'd like to see what kind of difference it would make over a stock GTX.

 

[flexy]

Originally posted by: BFG10K
Based on those benchmarks my next card will be a 7800 GTX. Of course a minor miracle could happen and the benchmarks could show something different in a month's time.

i am pretty happy with my X850XT@PE right now....and the next time when i get a new card will be when there's a game i NEED a new card to play it with.

Right now i can play D3/HL2 and ALL other available games with max settings/AA etc... on - SMOOTHLY and fast.

The time when we get a game/engine which demands more and NEEDS SM3.0 and all that (IN THE FUTURE) i guess R580 will be out and G80 and whatsoever.

It would be a totally different story if i had a low-end card right now and would really badly look forward to R520/7800GTX to be able to play games. If those benchies would be true (which i dont know) i would be VERY disappointed.

Also...even if those benchies are not true and the R520 would be about 5x as fast as the X850XT PE.....it's kind of an irony because i have no frickin' clue how to utilize such power because (as said) the game-engine which DEMANDS this is not out yet.

Let's wait for Vista (shudder, ehrm...)...the new unreal engine or whatever is coming to us...and while i am still happy with the X850XT i can just sit back, play games and watch prices for the newer cards fall and then grab one when i think i need it

 

[Banzai042]

Yea, i'm thinkin that i'm content to sit this round out, i've got a 6800Ultra AGP, so i'm not exactly hurting for graphics power at the moment. I figure i'll wait for the G80/R600 parts next summer and build a new gaming rig from the ground up.

 

[ronnn]

Originally posted by: keysplayr2003

Originally posted by: ronnn
These numbers are quite encouraging, as it has been expected for months that the r520 would be slower than the gtx. For early drivers, this is closer than I expected, so time to see what other features are in store. Good news for all that it looks like some competition is on the way. Or I suppose this article could turn out to be a fabrication and the r520 is even slower.

I'm impressed ronnn. It looks like you're the only one in this thread that isn't dissappointed. Kudos. Nice to see someone look for those silver linings. Kind of like KeepItRed always does.

Guess I should be insulted, but I'm not. :beer: I mean we all knew Ati would likely be slower this gen, but possibly richer in features - so time for us all to change sides, sort of like half-time. We all have a little rest and than the IQ wars start. (Is hdr really improved?)

Edit: Besides I saw no need to piss all over this thread, I mean supposedly the guy who released these numbers has a good reputation.

 

[Acanthus]

Originally posted by: Frackal

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

Actually with a voltmod and dual slot, I doubt it would be that difficult.

My GTX reports as 515 max, 75mhz higher isn't much, and I'm only 40mhz away from 1400, stock volts, stock cooling. IIRC a guy with 615/1450 got 11k in '05

Thats with cascade cooling on the card... A little different.

 

[Acanthus]

Originally posted by: Elfear

Originally posted by: munky

Originally posted by: Frackal

Originally posted by: munky

Originally posted by: ddogg

Originally posted by: Therk
I am actually very suprised of how well a 16 piper does against a 24 piper. I would love to see a 24 piper R520 against a GTX

look at the difference in clockspeeds between the two cards....theres almost 200mhz between the two and the GTX is factory clocked at 1200MHz on the memory which can be bumped easily to 1300mhz. i would love to see a GTX clocked at 600/1400 against this.

LOL, I'd love to see you try clocking a gtx to 600/1400, even with a voltmod and watercooling. Not gonna happen, just like the A64 wont reach 4ghz no matter what.

Actually with a voltmod and dual slot, I doubt it would be that difficult.

My GTX reports as 515 max, 75mhz higher isn't much, and I'm only 40mhz away from 1400, stock volts, stock cooling. IIRC a guy with 615/1450 got 11k in '05

You have to look at the percentage increase. 75 mhz on a 3 ghz cpu is not much. 75 mhz on a gpu that's already pushing the limit at 515 is a pretty big jump, actually.

edit: and that's 85 actually, 515 + 85 = 600

do you have a link to the guy who got his to 615?

Here you go munky. Lots of guys on the orb are running some crazy numbers. (Hit "Show me the projects", I guess it doesn't populate the results from the link)

ViperJohn, the volt-modding guru, told me that a his air-cooled cards get anywhere between 540-600Mhz on the core and one of his watercooled cards ran at 640Mhz artifact free.

I wish one of the guys with one of these uber cards would run some gaming benchmarks. I'd like to see what kind of difference it would make over a stock GTX.

So youre suggesting that a 7800GTX, which runs 70c load easy at stock, will be safe for long term use with a volt mod and OC that would dramatically increase that sustained heat.

 

[StrangerGuy]

Firstly, can we actually trust the HA benchmarks? Since no major reviewers such as AT, Techreport and Xbitlabs have released their R520 benchies yet , I feel we should take HA with a grain of salt for now.

 

[Keysplayr]

Originally posted by: StrangerGuy
Firstly, can we actually trust the HA benchmarks? Since no major reviewers such as AT, Techreport and Xbitlabs have released their R520 benchies yet , I feel we should take HA with a grain of salt for now.

I am doubting that AT, Techreport and Xbitlabs have these cards yet. Neither did HA but coaxed benches from a OEM partner (not named for good reason).
From what I hear, this Sander guy is reputable. So you probably can take these results with a salt shaker if you want, but I would not go as far as the salt flats.

There is always room for driver improvements I suppose. It depends on ATI codewarriors.

 

[ddogg]

whether you take those results with a pinch of salt or not....its basically just an overclocked X850XT with tweaks in the architecture, you cant expect this card to double the performance. those results seem pretty accurate for a card thats just basically Oced.

 

[xtknight]

Originally posted by: keysplayr2003
I am doubting that AT, Techreport and Xbitlabs have these cards yet. Neither did HA but coaxed benches from a OEM partner (not named for good reason).
From what I hear, this Sander guy is reputable. So you probably can take these results with a salt shaker if you want, but I would not go as far as the salt flats.

From what I hear, this is the same douche saying you needed a 5ghz p4 to beat an fx-55 or something along those lines. He doesn't paint a professional picture of himself. '7800GTX is ****ing fast' is one of his articles.

I'd believe anything I saw from HKEPC (Chinese site) though. They have been damn reliable. At this point I trust the inq more than Sander Sassen.

 

[KeepItRed]

Lol what if Sander made those benchies himself in Photoshop or something. Then made it sound real like ATI is commenting on it.

 

[nRollo]

Originally posted by: KeepItRed
Lol what if Sander made those benchies himself in Photoshop or something. Then made it sound real like ATI is commenting on it.

LOL what if he didn't and they are what he says they are?

 

[nitromullet]

Originally posted by: xtknight
From what I hear, this is the same douche saying you needed a 5ghz p4 to beat an fx-55 or something along those lines. He doesn't paint a professional picture of himself. '7800GTX is ****ing fast' is one of his articles.

I'd believe anything I saw from HKEPC (Chinese site) though. They have been damn reliable. At this point I trust the inq more than Sander Sassen.

http://www.hardwareanalysis.com/content/article/1787/

Why don't you read the article before calling him a douche?

Edit: I also don't see the article the 7800GTX is ****ing fast article you referred to, maybe you could post a link?

 

[nRollo]

Originally posted by: nitromullet

Originally posted by: xtknight
From what I hear, this is the same douche saying you needed a 5ghz p4 to beat an fx-55 or something along those lines. He doesn't paint a professional picture of himself. '7800GTX is ****ing fast' is one of his articles.

I'd believe anything I saw from HKEPC (Chinese site) though. They have been damn reliable. At this point I trust the inq more than Sander Sassen.

http://www.hardwareanalysis.com/content/article/1787/

Why don't you read the article before calling him a douche?

Edit: I also don't see the article the 7800GTX is ****ing fast article you referred to, maybe you could post a link?

Looks like a reasonable article to me? I think we all know Intel CPUs are a lot slower than AMD for gaming these days, he just checked out what it would take to put them on par.

 

[xtknight]

Originally posted by: nitromullet

http://www.hardwareanalysis.com/content/article/1787/

Why don't you read the article before calling him a douche?

I don't trust the numbers until other sites do it and come out with the same results. There is another article he did I'm trying to find at the moment, but it was pretty stupid. It wasn't the P4 one.

Edit: I also don't see the article the 7800GTX is ****ing fast article you referred to, maybe you could post a link?

http://www.hardwareanalysis.com/content/columns/article/1804/

 

[xtknight]

It was this one: 'AMD chipsets quirky by nature'

http://www.hardwareanalysis.com/content/columns/article/1786/

I'd love to see all these issues he talking about.

Edit: all I'm saying is I wouldn't trust him in a million years, just like some people don't trust the inq in a million years.

 

[nitromullet]

Ah, ok, you're looking under the 'Columns' section. Did you the read f*ing fast commentary? Just a commentary on NV marketing slogan, nothing very unprofessional or offensive about it.

I also tend to agree with him on the AMD chipset issue as well, as it reflects my own experiences. The rig I have now is great, but it's done some goofy stuff in it's time with me already and has taken a bit tweaking to get truely stable. If Intel can get their performance up my next rig will probably be Intel based, actually maybe a mac that will dual boot OSX/Windows I really don't want to get into Intel vs. AMD in this thread, and for the record I am very happy with my current set up.

Of course neither of these articles are really relevant to the benchmark article, since he admits that he didn't bench the card himself. Whether he's professional or not, one doesn't get the impression that he's a cronic BS'er... If his numbers are false, I don't think that it is because of Sander, but his source could be wrong.

Edit: I'm sure you can find a few chipset issues in this 209 page SLI FAQ thread... http://www.hardwareanalysis.com/content/topic/38623/