• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The first "proper" Kepler news Fri 17th Feb?

When you look at it that way, yes. When you look at "top dawg" prices there's no difference.

ATI has only been cheaper in the past because their products were slower.
 
Sorry Guru but you're wrong, sometimes they have charged Nvidia prices, then in the case of the 5870 and the 9700 PRO, where they were undisputed kings, they didn't
 
5870 was not the undisputed king though.

And the 9870 pro is the only example I can think of where ATI did not charge top dollar for their top dog.

Remember the X1800XT and XT1900 days? MUCHO expensive which was why I had a 7800 and 7900gt at the time.

Y U NO THINK IM RIGHT
 
tumblr_l96b01l36p1qdhmifo1_500.jpg
 
See what this thread has degraded into. Simply because Nvidia have failed to whet our appetites? ;)
 
I love the way the rumours are if anything stepping up and becoming even cloudier the closer we get to real information coming out - tomshardware today have reposted the pics of the new card from xbit but not read the article on xbit so they are still saying the GK104 is GTX660, where as most other sources (including the one they are quoting) are saying it's now going to be released at GTX680
 
the ram chips on current 7970 are rated to 6ghz anyway - they just dont run at that *yet*

Hynix H5GQ2H24MFR are the chips - but its down to the controller on the chip
 
Harlequin said:
the ram chips on current 7970 are rated to 6ghz anyway - they just dont run at that *yet*

Hynix H5GQ2H24MFR are the chips - but its down to the controller on the chip
Click to expand...

Not even that, its down to what's required, if it had a 256bit bus it would have overclocked every single last Mhz out of the memory and used more power doing so. With the 384bit bus there is just so much ruddy bandwidth spare there isn't any need to run the memory at 6Ghz, when you can run lower, lower voltage, lower voltage on the gpu core because the memory core can run lower voltage, etc, etc.

This is why the rumoured very low core clocks with as high as possible memory clocks are very suggestive of where the performance limit will be on the GK104. You save power where you simply don't need to spend that power, memory clocked to high hell and core clocks well below last gen, on a new process suggests the same power saving AMD is doing on the memory, but for Nvidia they are doing it on the CPU.
 
Broken Hope said:
http://hardforum.com/showpost.php?p=1038464236&postcount=51
Click to expand...

The irony. Now AMD might do better in tessellation than Nvidia within the same generation! :p

I wonder how much faster the GTX680 will be?? Would an HD7980 or pre-overclocked HD7970 be able to catch it??

How much will it cost?? The lowest priced high end Nvidia card was the GTX285 for around £295 to £350.

So,I suspect at least £300 but it could be around £400 to £450.
 
I don't care who does what, just bring a new high end card to market soon. I need something to play games at 2560x1440, my venerable 5970 doesn't quite cut it any more, nor does a single 7970. We need moar power to the engins captain...
 
http://semiaccurate.com/2012/03/07/tsmc-suddenly-halts-28nm-production/

Sit down before you read this one, we will wait for a second while you do. OK, now, where to begin other than the obvious, TSMC shut down 28nm production about three weeks ago.

No, we are not joking, word has reached SemiAccurate that TSMC halted 100% of 28nm production in mid-February to make unnamed changes to the process. Since only Nvidia was having issues, and those are not really process related, we have no explanation as to why they would take this drastic step, but yields don’t seem to be it. That said, we are quite certain that it happened, and is still in effect.

Sources tell SemiAccurate that the stoppage was quite sudden, and production should resume in short order. One source said that by the end of March, volumes would be back to where they were before if not a lot better. Checks with the channel have confirmed an abrupt stop in 28nm chip shipments, not that there were many to begin with. TSMC is privately promising a quick and planned down time with a resumption in production very soon.

Whether wafers in process were continued, paused, or scrapped is not known. If the parts that were started can be continued on once the switches are flipped back to ‘go’, the production blip won’t be that bad. Because all are saying the whole episode will be but a bad memory by month’s end, we believe that wafers were just paused. If that were not the case, if production restarted today, you would not be seeing new chips coming out until June.

Checks with various points in the chain for graphics cards show that there was actually a large amount of supply of 28nm AMD parts in the channel from chips to cards. That supply, while not infinite, should be enough to keep products on the shelves until new parts come back out. Basically channel inventory will be depleted about the time it starts being replenished. Best case, buyers won’t see any problems, worst case, 1-2 weeks of short supply. AIBs however have to sit on their thumbs and do nothing until the new parts come in, then it is overtime shifts until demand is sated.

Things get much more complex for Qualcomm. 28nm Kraits are not shipping to end users, and as far as we have heard, nothing has been given a definitive ship date. If there is a 2-3 week delay, the OEMs and carriers will be peeved, but the end user will have no clue about any problem. It is internal, logistical, and annoying, but not a PR disaster.

Nvidia on the other hand is announcing their newest family of parts next Monday. It is a good part, but supply is the proverbial elephant in the room. Ivy Bridge is launching soon, and if GPUs aren’t ready, things get ugly in a hurry. Given Nvidia’s uncharacteristic whining during the last conference call, you now understand what they were prepping the analyst community for. SemiAccurate still believes that Nvidia is the only company seeing 28nm ‘yield problems’ and the real problem has it’s roots in Santa Clara.

All in all, this is nothing more than a horrific failure. Complete 28nm stoppage at TSMC with no public word, no explanation, and no known concrete resumption date. Sources tell us that it is indeed temporary, but until there is a good explanation, don’t assume anything.S|A
Click to expand...
 
Sampsy said:
It really does seem that all semiconductor foundries other than intel's own are completely incompetant.
Click to expand...

That's a bit harsh.

Intel's own foundries are tailored specifically to Intel's needs. All other foundries must accommodate a wide range of clients, who are likely to have radically different requirements. What works well for one application may not be so effective for others.


Regarding the story above - that could be quite a setback for those companies relying on 28nm TSMC chips. We have no idea what the reason is, but we shouldn't pre-judge - it could simply be a shut-down in order to upgrade the throughput (which would be a good thing in the longer term).
 
You must log in or register to reply here.
Back
Top Bottom