• 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.

16nm GM200 Titan II being tested..

JediFragger said:
DM's posts are at least informative, even if you don't agree with them.

Suggest you stop trolling now, making yourself look like a bit of a moron tbh
Click to expand...

Not really, DM's posts are still all based on rumour and supposition, they are no more valid than the person he was berating

No need to start throwing insults about - you or him - perhaps take a look at your own posts if you have to resort to that
 
drunkenmaster said:
I do love the generation of people that think anything over 140 letters is a wall of text. The response to posts using more than a few words does often identify people who really aren't worth talking to.

Care to discuss any point I made at all, if not, honestly, why are you even here? Are you so bored in life that you go on a DISCUSSION FORUM to get upset and people trying to discuss things?


For the record I was responding to someone who said that it was sure as hell not a 16nm product.

http://www.xbitlabs.com/news/other/..._Regarding_16nm_FinFET_and_20nm_Progress.html

Summed up, October 2013 TSMC has taped out multiple 20 and 16nm chips.

http://techreport.com/news/26078/arm-and-tsmc-tape-out-test-chip-on-16-nm-finfet-process

AMD/ARM tape out complex test chips late 2013. 16nm products were taping out as long as a year ago, it's entirely possible for Nvidia to have early 16nm chip samples. it isn't an indication of TSMC being ready to ship in volume or as I said, that the chip is actually capable of being produced in volume, it may not work, it may have bugs or it may have yield issues which could all necessitate a respin.

Someone said it couldn't be 16nm, I pointed out it could be, discussion... on a discussion forum. I also added reasons and explanations of why a chip turning up in August doesn't mean it's automatically ready to ship as a final product. Such a chip shipment would have happened for the A1 silicon of the 480gtx, 4-5 months before they got the respun version of it back as A2/B1 silicon. A chip existing does not mean it's certain to be a product.
Click to expand...

Well actually for the record you were replying to someone that was quoting from the article the OP linked that suggested it wouldn't be 16nm when he said it would.
 
andybird123 said:
Not really, DM's posts are still all based on rumour and supposition, they are no more valid than the person he was berating

No need to start throwing insults about - you or him - perhaps take a look at your own posts if you have to resort to that
Click to expand...

My posts are based on rumours? Even though I linked to two articles where TSMC directly stated multiple 16nm products had taped out and ARM stated they have taped out a product test arm chips(this is really all ARM actually do because they create effectively working based blocks that then other companies can use and add to, ARM don't tape out full commercial chips on their own). Learn the difference between a rumour mill site like wccf and a story that used a piece of shipping information with no direct information on the process, the part, if it's working, if it's a release ready silicon and imply whatever they hell they want and a credible site writing an article based around press releases direct from companies like TSMC/ARM who confirmed such tape outs. One is reliable and the source can be found to be TSMC/ARM, the other is click bait rubbish that made no attempt to suggest why it was impossible for a test chip to reach Nvidia in August that was made on 16nm even though 16nm chips taped out at least 10 months before it. They simply claimed it was impossible with no explanation, largely because there isn't one.

I wasn't berating anyone, I just explained why the "they sure as hell aren't 16nm" part was false.

This is the problem with in particular you and several others. Show me where i berated the guy, differing in opinion is not berating. Using more words is not berating. My post was discussion, nothing more or less, but everyone time someone makes a point you jump to the position of I was berating someone. I do question why people are on discussion forums if they never want to discuss anything at all.

In yet another thread your only contribution is to have a go at certain people and have added nothing meaningful to the discussion in the slightest.
 
Last edited:
triss said:
Well actually for the record you were replying to someone that was quoting from the article the OP linked that suggested it wouldn't be 16nm when he said it would.
Click to expand...

For the record you included nothing in your post that showed you were directly quoting from said article, most people would actually put it in quotes, or use the quote tags around it, or actually say at some point you were quoting the article, linking to the article doesn't mean you're quoting from it.

Anyway as such, nothing changes, the article is also wrong. They have zero proof and absolutely no first hand information that the product "sure as hell isn't 16nm". They are going on... it was in August thus must be 28nm.

I'm not and haven't said it is 16nm or can't be 28nm, simple that stating it couldn't be is incorrect. This chip was shipped to Nvidia around 10 months after multiple other 16nm tape outs were confirmed.
 
andybird123 said:
Not really, DM's posts are still all based on rumour and supposition, they are no more valid than the person he was berating

No need to start throwing insults about - you or him - perhaps take a look at your own posts if you have to resort to that
Click to expand...

Well I (and I suspect a lot of others) appreciate DM's posts, wall of text and all! :D

If people can't handle that amount of reading they really should just stick to Twitter ;)
 
Last edited:
triss said:
Well, the consumer samples taped out a long time back, and they are sure as hell not on 16nm FinFET

Read more: http://wccftech.com/article/generat...a-big-daddy-maxwell-16ff-ports/#ixzz3IepiaNq4

From the article you linked
Click to expand...

drunkenmaster said:
For the record you included nothing in your post that showed you were directly quoting from said article, most people would actually put it in quotes, or use the quote tags around it, or actually say at some point you were quoting the article, linking to the article doesn't mean you're quoting from it.

Anyway as such, nothing changes, the article is also wrong. They have zero proof and absolutely no first hand information that the product "sure as hell isn't 16nm". They are going on... it was in August thus must be 28nm.

I'm not and haven't said it is 16nm or can't be 28nm, simple that stating it couldn't be is incorrect. This chip was shipped to Nvidia around 10 months after multiple other 16nm tape outs were confirmed.
Click to expand...

Yes i didnt use quotes but it was mentioned .
Sorry for not using the correct form, but maybe if you had read it and even the article linked there and in the OP you would have seen it indeed it was in the link rather then just trying to flame someone
 
Orangey said:
384 is all you get.
Click to expand...

3GB or 6GB then as I doubt they would use any trickery to a flagship range to add mismatched VRam buses like they did on lower cards in the past.

The R&D was done long ago (excuse they used in past was R&D cost) for the 512bit bus as the GTX 280 and GTX 285 had 512.

I know it matters less to GDDR5 over GDDR3 as its QuadPumped but things need progress.
 
jh30uk said:
3GB or 6GB then as I doubt they would use any trickery to a flagship range to add mismatched VRam buses like they did on lower cards in the past.

The R&D was done long ago (excuse they used in past was R&D cost) for the 512bit bus as the GTX 280 and GTX 285 had 512.

I know it matters less to GDDR5 over GDDR3 as its QuadPumped but things need progress.
Click to expand...

Would 9GB also work ?
 
triss said:
Would 9GB also work ?
Click to expand...


Not sure as the next workstation card up the range was 12GB up from 6GB, depends on the modules size in relation to each SMX (or is it the ROPS) as that determines how many memory controlers it can have so how many modules per controller.

I cannot remember the terms correctly but someone like Roff will know.
 
Last edited:
triss said:
Yes i didnt use quotes but it was mentioned .
Sorry for not using the correct form, but maybe if you had read it and even the article linked there and in the OP you would have seen it indeed it was in the link rather then just trying to flame someone
Click to expand...

Probably best just to add him to your ignore list tbh, that way his trash only shows in quotes.
 
The interesting thing is that card is 50% more than 980 in every way. 50% more shaders, 50% more bus width, etc.

So you can have a good guess about performance :)
 
You must log in or register to reply here.
Back
Top Bottom