Yeah, now that would be cool since HD 2600 seems to be the card that I can probably afford:p:
Printable View
Yeah, now that would be cool since HD 2600 seems to be the card that I can probably afford:p:
I wouldnt say they are at their peak. yes a new design is needed and im sure nvidia and ati have new architecture ideas on paper.the way you say it sounds like it will be the end of new GPUs in one more generation. I dont think so, or you need to reword your statement.
That's mute when there are dozens of DX9 games and only 1-2 DX10 games :slap:
Besides, if you read this it's becoming apparent they have a very old version of the R600. Remember you are only looking at a xtx that's clocked 50 higher then the XT. Even if the retail version of the XTX is at 800/2200 that's still higher then what 8800 is set to and still shows higher performance (from the GTX I am not inclined to say the same with the GTS do to problems read in posts).
Don't forget that it's a good sign that the xt/xtx retail version my crush the 8800. This IMO as the result of lifting NDA next week instead of the 14th. In other words, when you have a good product, there is no need to hide it from the competition <barcelona> COUGH.
http://forum.beyond3d.com/showpost.p...postcount=3126Quote:
Originally Posted by Kombatant
Hasn't anyone learned anything from CPU's? Higher clocks doesn't mean its faster than the competitor.
Sure its not a great performer in those benches but well now that I think of it. Surely the 8800GTS was an OCed card as well. Wish they showed some OC benches on the 2900XT.
death, unfortunately it's actually becoming increasingly difficult to find reference clocked NVidia cards.
Going by the definition of "stock clocks" though, the out-of-the-box clock speeds are stock. PNY are one of the only brands that still sell reference clocked NV cards.
Now, if ATi allow vendors to sell overclocked cards, then you may have a valid argument. For now though, the fact that you can buy the 8 series in stores at those clocks makes them stock.
PNY, EVGA, Leadtek, Zogis, Asus, XFX. Those are all the manufactures that have original manufacture clocked cards at newegg. So just more than one manufacture sells reference cards and that was just the 8800GTS series (640 and 320meg).
http://www.newegg.com/Product/Produc...page=1&bop=And
Guess I should of said reference instead of stock when I was talking. Question is where they using a reference board or a faster one.
Sorry but that's not correct. The stock clocks can always be found at Nvidia's homepage. You don't go by what some 3rd party vendor decides to OC it to then declare that's stock settings. Those have always been and will always be OC settings "supposedly" guaranteed by that vendor.
............................8800GTX....8800GTS.... 8600GTS.....8600GT....8500GT
Core Clock (MHz)..........575..........500..........675...... ....540..........450
Shader Clock (MHz)......1350........1200........1450.........11 90.........900
Memory Clock (MHz)......900..........800........1000..........7 00..........400
You do have a valid point. They need one hell of a marketing plan based on the right price points and they can be successful with this card. Pricing it in-line with the 8800s would be a bad idea, so I somehow doubt they will be doing that, unless they are prepared for a serious uphill battle.
All my ati cards were stellar overclockers...
Have you ever seen a 12 pipe X800pro (not flashed to 16 pipes) break 7k in 3dmk05 on air?
How about a non flashed 9800pro clearing 4k in 05?
Sorry u had bad luck...but its probably just the samples u got, not ATI. :(
They are one and the same in this example (in the context that it's been used). Unless you can provide some core arch. GPU changes that make one different from another? If all you are referencing is the "clock speed" then it's simply the same. Anyone can change the clock speed of a video card. But it doesn't make it "the new" stock clock for that video card.
being honest im about tired of all the speculation so far.
bring on the day when the card is out there for us all to get our hands on and get something that isnt pretty much just hear say and 'news'.
okay a student like me wont be able to afford one but at least the information around the web will likely be a bit more accurate, what with all the 'old version of the cards' and 'old drivers'.
i just cant wait till there are full reviews and not just small table with a couple of numbers on.
there's a difference between speculation and news.
I have had pretty much the same success with my ATI cards. It is because of this that I no longer get lower models in hopes of reaching higher model speeds after an overclock. It just never seems to work out that way. Starting from my X1900XTX I am always only buying the very top end part and hoping for a minimal overclock, that way I am still on top an not having buyers remorse as $100 to me is worth knowing I have the fastest, and not regretting it a week later that the $100 I saved could have been made up with my next pay cheque, yet I am stuck with what I feel is an under performing part than what I could have had.
I'm with Diltech on this. I do see the other posters side loud and clear, but to me stock clocks are what card is suppose to be doing out of the box as set by the manufacturer. Reference clocks to me are clocks set as a recommendation or a suggestion from the chip designer that the card should be run at.
The manufacturer considers the clock speed the card is sold at as it's stock speed. They warranty the card to run at that speed, right? They sell it running at that speed, with no modifications, be it software or hardware, required to make it run that speed right?
Wouldn't that technically, by any means of the term, make it the cards stock speed?
A long good read: Tech ARPQuote:
The ATI R6xx And What It Means To Us
Although this may seem to be the start of some mindless bashing of AMD and ATI, please don't judge me so harshly just yet. This article is not about how AMD and ATI screwed up. It's about what the R600 means to us, the average computer enthusiast; and what will the VPU affect AMD-ATI's future? With the launch fast approaching, it would be good to look at what ATI fans can look forward to in the R600.
(Stock speed to me it was it runs at before overclocking)
Please remember that DAAMIT has every reason to make their upcoming products look like crap (hint: 8800U, 8600GTX, etc). I'm pretty convinced that the retail products will perform a lot better.
No, because what you suggest makes it no different from what any user with ATI tool, ATI Tray, Rivatuner, etc can do on their own. Also, any user can alter the bios to change the clock speed. This alone validates my point in this example. Just because a vendor finds a good OC doesn't make it a new stock clock speed. It only makes it the new over clocked speed. There is no reason to re-invent over clocking into something that it is not. Or else we repeat what we should have learned from the past .
well, at least u can fold on it.
i hope there is a HD2950 or 8900GTX thats single card (not like 7950 GX2) .... or if nv do get 9800 out b4 end of this year ill get that , too...
It seems like DT numbers are a bit off
Link:Quote:
Bum_JCRules wrote:
While I am required to follow the NDA, the stuff up on Daily Tech today is almost worthless. Yes Anandtech was present in Tunisia (signing Non-disclosure agreements like the Inquirer), why they are posting this stuff is beyond me because their numbers are off. They must be only using the XP drivers and OS because the numbers in CF vs the GTX are very much different. So until I can officially comment on the architecture and the performance.. hold all of this as useless until the rest of the world writes about it.
http://forumz.tomshardware.com/hardw...derasc-75.html
Also Kristopher Kubicki in his blog mentions new drivers and CF setup:
Link:Quote:
I was curious about the lack of HDMI as well, though one partner told me on the XT the only HDMI capabilities would be via a dongle (so no sound obviously).
Someone "reached out and touched us" today and gave us 2 Radeon XT cards for use in Crossfire and a different driver (350MB ?!?!). Expect benchmarks this week.
http://www.dailytech.com/DailyTech+D...rticle7079.htm
Lets wait and see ;)
Hector Ruiz wouldn't allow a catastrophe of that magnitude...Dave Orton would...but Hector wouldn't...you can't even think about fighting a multi-billion dollar monopoly if you take out products 8 months later only to be inferior to the competitor.
The $399 HD 2900 XT will kick the 8800's behind all the way from Tunisia to Santa Clara.
Perkam
Exactly lol, to be honest I think you should do teh same as me and wait a month or 2 after it's released to ensure a good sample. I do not, and can not, deal with another poor overclocker. My CPU already pisses me off because of that fact, it's one of the first ones as well typically...
Hector wouldn't allow a catastrophe of that magnitude? LOL Hector can't even avoid creating a catastrophe of that magnitude all by himself. Bigger in fact, because that exactly what the ATI acquisition has turned into. He's now mortgaged AMD's future under a pile of debt the likes of which AMD has never faced before. As for his ideas about how to fight a multi-billion dollar monopoly, did you listen to AMD's last conference call and all of Hector's blathering about going "asset-light"? Asset light and debt heavy? Clueless and laughable if he thinks AMD is going to be able to take on Intel like that.
Hector is an idiot who's only successes were 100% the result of the efforts of his predecessor Jerry Sanders. The sooner AMD gets rid of him, the better off AMD will be.
It isn't a matter of allowing or not, if the engineers can't get results from the path they've taken and cannot find a better one its not directly the CEOs fault. There are fewer industries more tedious, and even those in far more straightforward lines of business blatantly screw up with end products, even while under great pressure to compete.
Nvidia had far less to excuse itself for the NV30 line than ATI has if the R600 doesnt turn out great, and that didn't stop the FX from happening.
Shouldn't it have higher performance scores in XP, considering how Vista is still far from gaming-ready? If they were benching XP 8800 against Vista 2900 then that statement might make more sense.Quote:
They must be only using the XP drivers and OS because the numbers in CF vs the GTX are very much different.
Also...CF numbers are different against the GTX? Double card is different against a single card, really?
It was. The direction they took was deliberate and completely at the feet of management. They pointificated about their vision over a long period of time in many interviews and conferences. "not about core anymore" being a phrase Hector would repeatedly say and Meyer parroted. It was going to be about platform and integrated video, and they described how core improvements, clock rate, and number of cores, ie cpu performance, wasn't important anymore. The only problem was they were wrong. Given c2d as an option, people bought those instead of "good enough"/"let them eat cake" AMD chips, which resulted in their losses. Taking heed of reality slapping them in the face, they've changed course somewhat, which has resulted in more improvements(claims of) in the k10.
Well in this case it was the CEOs fault originally.
Was just saying that CEOs cannot simply steer a company from destruction if their entire engineering are putting their all into a ground up new design and it isn't measuring up to the competition. You can't order someone to make something better if they honestly don't know how to with the given resources.
http://r800.blogspot.com/
Said to be test slides from AMD/ATI. XT vs GTS.
This scaling is more than amazing!!
If this is true then :woot: and my previous post makes it very probably it is!!
I wonder how HD2900XT CF will compete with nV8800GTX SLI then ...
PS. Now I know why 8800GTX is faster in most todays games ...
Attachment 58119
16 texture units of ATI/AMD R600 versus 32 texture units of nVidia G80 GTX makes a difference.
http://bp2.blogger.com/_BabjUDZIqPw/...c/s400/004.jpg
This is going for my desktop wallpaper.
http://i4.photobucket.com/albums/y11...o/gpun2076.jpg
2D clocks are 507/514, and yet it idles at 68 degrees C. Lets hope they took that pic in Africa. :)
Stock coolers are for nubs.
Comparing an enthusiast card to a mainstream-end one. Yeah, that doesn't look suspicious at all. :slap:
Dailytechs benchmarks are inaccurate. Atleast in terms of 3dmark 06 ;)
There are more leaked benchmarks that just came out that expalin alot IMO. I believe they are accurate and claim only 16TMU's, this would explain how it is comparable to the GTS. This also reminds me of the x1800, shortly followed by a more powerful x1900.
They probably want to start off with the 16TMU version, then will have the refresh with more.
2900XT is a little slower than 8800GTX
I'm sorry, but I don't think you know what you're talking about. One of the most 'asset light' companies in the world (in proportion to gross revenue) is wal-mart. You're telling me thier 'just in time' inventory system is a bad idea? Isn't that what revolutionized the japanese car industry in the mid 20th century? This type of inventory is incredibly efficient and can save companies a lot of money. Business students are taught that inventory is bad, and it is. If you have it, you may as well have lost the money until it is sold. It's heavy, it depreciates, and it clogs up the business structure.
He's talking about inventory systems and trying to not sit on chips that depreciate every minute they're in their posession.
Ryan
Heh. Obviously you didn't listen to the conference call, and have no idea whatsoever what you are talking about. Here is a link to the transcript of the conference call.
http://seekingalpha.com/article/32901
Here are a bunch of quotes from Hector himself during that CC making it blindingly obvious what he is talking about:
Quote:
As part of that, one of the things that has become pretty clear in our experience, we have had now for a number of years had some experience in partnering with people such as, for example, IBM in joint development programs. That has started a lot about how you can do some asset light strategies, since we did not have to build an R&D laboratory to do that.
We have had also an experience now for a number of years with some of our friends in the foundry business, and in particular the charter semiconductor. We have learned a lot of that.
Through the acquisition of ATI, we now have a perspective into a very asset light model that we are pretty excited about learning more and more about it. When we look at all these things, we see a tremendous opportunity for us to really do something different going forward, but it’s unique for us.
I think, for example, what Texas Instruments has done, as you said, is very good for them and is uniquely tailored for what they do. Whatever we do is going to have to be very uniquely tailored for us, because we have to be on one hand, in the workstation super computer environment at the leading, bleeding edge of technology, but on the entry level part of products is not quite as necessary.
I think we are in a unique segment in the industry, the microprocessor industry where an asset light model may require a considerably different look and feel that you might recognize when you look at somebody like Texas Instruments.
Quote:
It coves a whole gamut of opportunities, from everything like I mentioned earlier, we already have a joint development agreement with IBM which we view that as an asset light strategy from the point of view that we do not have to build an R&D facility, and all the way in to any aspect of the rest of the work that we do.
Here he is talking about cutting down on capital investments and R&D through ousourcing. He mentioned the partnership with IBM (where they have been oursourcing their process R&D), and the partnership with Chartered (where they have been oursourcing their production), and talking about making themselves resemble Texas Instruments more (which just announced they were exiting process development and going to go totally fabless). Nothing whatsoever about inventories or anything you were talking about.Quote:
and reduce our capital intensity by exploring deeply more asset light business models in order to fully execute our plan.
Finnaly, it´s showtime....
Bring it on Kinc
any bets that the xtx will have 32 texture units
Even though I'm not getting one of these (or the 8800 either) I'm still impatient for it to come out. Maybe because I'm tired of all the speculation....
see http://r800.blogspot.com/2007/04/rad...2900xt_28.html
it looks like truth
these cards look very promising, crossfire mode is insane!!
Disappointed at the preliminary performance numbers, but can't complain about the price tag.
We certainly don't need more $500-600 video cards in the market.
Not only that but no info on the 8800GTS used & that site is about as reliable as a bucket with holes in it.
That's the sort of thing you'll find on that blog. Can we not ban the posting of these piece of crap fanboy blogs?Quote:
Congratulation! AMD ATI and all perseverance ATI fans in CHIPHELL and around the globe.
http://www.chiphell.com/attachments/...U82mucxbN6.jpg
http://forum.beyond3d.com/showpost.p...postcount=3293
R600XT 745/828, C2D E6600 @ 3.6GHz
http://www.chiphell.com/attachments/...cq13fv8Zhx.jpg
http://forum.beyond3d.com/showpost.p...postcount=3451
Look at the post by Kab @ 04/29/2007 03:56 AM
http://forums.amd.com/forum/messagev...VIEWTMP=Linear
I guess there is hope yet, good, cause if the floped now, i don't see Nvidia getting of their a$$es and releasing something new in the fall.
First of all, those are official benchmarks released by ATI. Not that they are worth anything to me or you.
Second of all, they used 8xAA/16xAF and adaptive AA. ATI's 8x AA implementation is different from nvidias, so Nvidia cards take a larger hit from that alone. ATI always had more efficient AA algorithms, so its no surprise its wins in most benchmarks.
Third point is, its compared to the 8800GTS. 8800GTS is not the 8800GTX so no great achievement there from ATI.
But like Kinc said, too many variables for us to take anyone's word for anything...
http://www.fudzilla.com/index.php?op...d=728&Itemid=1
So fuad gives us hope that this still maybe a g80 killer.....Quote:
R600 has 16 TMU's while G80 has 32
R600 has 16 TMU's while G80 has 32
Written by Fuad Abazovic
Saturday, 28 April 2007 10:08
That is why it stays behind
Here is the one of the key reasons why R600 marchitecture falls behind the G80 GTX version. While Geforce 8800 GTX or G80GTX has 32 Texture Memory Units (TMUs) Radeon HD 2900 XT has 16 only.
TMUs are responsible for getting the textures in your games and more you have the faster you can handle the texture work.
Still, the Radeon HD 2900 XT at 750 MHz core is about 30 percent faster in raw clock speed from Geforce 8800 GTX at 575 MHz and finishes the texture cycle some 30 percent faster then Nvidia.
In spite of that this is not really a fortunate solution as Nvidia has twice as much texture memory units. The only hope is that ATI has more effective TMUs, but this is one big marchitectural disadvantage over the G80 marchitecture.
One more thing, according to beyond3d, the latest release-candidate driver from ATI causes the r600 to do weird and wonderful things like destroy an 8800gtx in performance...
http://forum.beyond3d.com/showthread...39173&page=144
And some linkies:
http://rage3d.com/board/showthread.p...post1334917155
Time to get some sleep now.
http://www.xtremesystems.org/forums/...1&d=1177841274
this is the best hype i've ever seen, so many benchmark scores and ll completely different (one showing a slow card not keeping up with the gts, the other one a tie between GTX and XT and the third a huge lead for ati)
everyone seems to talk about their product, some bashing and some praying but no one rlly nows, this is better than cinema :D
BTW: pcgameshardware (germany) got no info about a launch on may 2nd, maybe they're releasing final drivers for testing on this date, NDA still seems to be lifted on 14th:
http://www.pcgameshardware.de/?article_id=600354
More details about R600:
http://img47.imageshack.us/img47/1470/flopsiw6.jpg
http://www.forum-3dcenter.org/vbulle...&postcount=351
http://img453.imageshack.us/img453/9275/21422560ju9.jpg
http://www.forum-3dcenter.org/vbulle...&postcount=364
R600 Pipeline Utilisation
http://www.cupidity.f9.co.uk/b3d95.gif
http://forum.beyond3d.com/showpost.p...postcount=3679
Only 60-80% faster in mosts games than the old 1950XTX.
http://www.directupload.net/images/070429/lhG7bb7V.jpg
http://www.directupload.net/images/070429/4Mv6EmTP.jpg
Maybe this will make for some killer price drops after all :) . I am all for killer price drops despite whoever has the crown.
On a Side note: I HATE BENCHMARK GRAPHS THAT DO NOT START AT 0 !!
Edited note: Those last graphs are for special circumstances only ! Notice they are with HDR and the second one is for stencil shadow, whatever that may be.
I can't wait to see some real benchmarks!!
I don't understand you.......
Inside joke or something about what I said? I meant how they started the graph with 80% making it seems further stretched out of a difference and also how it is using % and not actual FPS.
My 8800GTS maxes out my westinghouse 27" actually at 1365x768 in all games so far.
What part of "60-80%" warranted an "Only" in the beginning of the sentence ?Quote:
Only 60-80% faster in mosts games than the old 1950XTX.
If ppl keep that attitude up, there will be a lot of jaws dropping on launch day...
Perkam