RV870 Story: AMD Showing up to the Fight

Sad to hear about sideport, I was always holding out hope that they could have gotten around to enabling it somehow, maybe in a driver update or something, but if it's not in the newer designs that is obviously not going to be a priority anymore.
 
Good article.
I've always wondered whatever happened with [H] and AT. I remember events sponsored by [H] where Anand has given speeches and Kyle talking about Anand like a friend to where I don't even see review links to AT articles anymore. Was there a deviation which I missed?

Sorry, I know this is a bit off topic, but I hope it is not deleted.
 
Awesome article. This "behind closed doors" look at an IHV is something I would like to see more of. I would love to see what AMD and Intel have going on in the CPU department.
 
That was a wonderful article; they have some great journalists and content over there. Hopefully AMD is still on the upswing now; I'm definitely looking forward to "Northern Islands." :)
 
Great article. I love these "behind the scenes" looks because they're usually completely in secret

The RV740 section confirms a lot of the suspicions behind what's possibly keeping Fermi behind so long
 
I had to check the date. Anandtech has already written most of it in other articles.
Good read anyway. Always nice to get details.
 
I had to check the date. Anandtech has already written most of it in other articles.
Good read anyway. Always nice to get details.

Yeah that was strange, because I was sure I already read that month ago.
 
after reading that.. and seeing how the Cypress cores turned out.. makes me wonder just what ATI has up their sleeves with the Northern island cores.. going to be a fun year indeed.. i hope Nvidia's ready.. :D
 
That really was a fascinating article, it makes video cards not seem like just a newer faster product. The presence of that article gives me a huge respect for the risk and probability factors that go into video card creation. It was very interesting, and I'm glad that going to 40nm early was a bet worth the risk. I'm looking forward to that first generation Holodeck in 2016! Ha!
 
thanks for linking this article. i don't typically look at anand, but that was a really well-written and insightful read.
 
That really was a fascinating article, it makes video cards not seem like just a newer faster product. The presence of that article gives me a huge respect for the risk and probability factors that go into video card creation. It was very interesting, and I'm glad that going to 40nm early was a bet worth the risk. I'm looking forward to that first generation Holodeck in 2016! Ha!

It was crazy seeing that mentioned, but he has very high goals for ATI getting that far. :D

I'm eager now to see what "Northern Islands" is going to be. Nonetheless, that was a very interesting read plus a very well written article at went behind the RV870 GPUs.
 
This is a great article. Thanks for the link.

After reading it my thoughts are that if ATI can continue to execute as well as they have these past few years (and all signs point to that trend continuing) NVIDIA is in for some very rough times in the GPU business. They simply cannot afford to continue with business as usual.
 
this is the first article that makes me read for so long and enjoy so much..

I have no word to describe it beside "WOW"

I am really impress with this, and their strategy to get 5000 series out..

but too bad about sideport, I really hope this thing can pop up someday in future, very anticipate....
 
in for the holodeck in 2016, but i'll probably wait for the next generation hardware so i can play previous generation holodeck games at higher eye candy
 
Very good article. Was just going to skim it a little bit since it is 3am over here and ended up reading the whole thing.
 
Abolustly fantastic read into the mind of an ATI engineer. Writing was very interesting and for someone that loves tech, this was a real treat. Props all around.
 
It's a lot of good information and good insight into the world of GPUs. Perhaps it'll get someone to show some appreciation for the magic that is doubling preformance every 18-24 months. And by magic I mean some damn fine work and overtime by some engineers.
 
Great article. I can't think of a single other tech company that sits down with journalists and lays out the background between products so honestly. And this is the second time in a couple of years I've seen ATI/AMD folks do this--both times with Anand.
 
I forgot to comment about this earlier. But I was amazed at how stressful this job is.

Its VERY easy to forget just how absurdly difficult it is to go to a different process.. And hopefully this will keep people from expecting the 6x series in late 2010 instead of early 2011.
 
Excellent read. Thanks for the link, I don't get over to Anandtech often enough these days :)

I've got masters-level classes in chip design, and the one thing I've learned is this: I do NOT want to get into the chip design business. It is not the carefree place it was 20 years ago, or the garage operation it was 40 years ago. The recent trend in complex process changes have made the business risky, caused smaller players to go-under, and have taken all the fun out of design.
 
Excellent read. Thanks for the link, I don't get over to Anandtech often enough these days :)

I've got masters-level classes in chip design, and the one thing I've learned is this: I do NOT want to get into the chip design business. It is not the carefree place it was 20 years ago, or the garage operation it was 40 years ago. The recent trend in complex process changes have made the business risky, caused smaller players to go-under, and have taken all the fun out of design.

After reading the article I saw that as well. The article stated that in 6 months, 3 engineers left to join nvidia. From what i can tell from the article, their new rules for getting stuff out made the whole thing a hard and grueling process, which is why people left. (least, that's how i've perceived it.) Nvidia may be taking their time with Fermi, but i've seen ups and downs for both sides.

After having built 5 gaming systems in the past couple months, the last one i built had a 5770. Although i'll agree, the 5xxx series really does perform impressively, the drivers i found lacked. I have a friend that 38xx and he always seemed to have problems gaming with it. We're going to be building a new for him at the end of the week and i don't want to be stuck with getting him 5770 but all the nvidia cards are getting less and less available so i won't have much choice.

My own rig has a 250 and my roommates has 260 and i haven't had a single issue with either card. Even the drivers off the cd it came with played all the games we threw at it.
The 5770 put into the one rig artifacted in a couple games. I updated to the latest drivers and the issues were gone. Yes, i know i should have done that to begin with, but it's interesting to do these things. In all my experiences, each nvidia card i've seen has worked right out of the box and the driver is fine for a good while before really needing and update. Which is the way i like it cause we all know, the customer doesn't care about anything other then it working and staying that way. Who wants to update their drivers every time a new game comes out?

All in all though, that was great read.
 
Man, reading articles by anand always makes me feel like I've gotten smarter. Even if I didnt. :D

The money quote, about Eyefinity:

"His desire to do this wasn’t born out of pure lunacy, Carrell does have a goal in mind. Within the next 6 years he wants to have a first generation holodeck operational. A first generation holodeck would be composed of a 180 degree hemispherical display with both positionally and phase accurate sound."


All the secrecy around Eyefinity makes it sound like they got the jump not only on Nvidia, but their own company and the game vendors as well. Calling it a different codename with each OEM to trace leaks, now thats planning. A bit early for a post-mortem on the 5xxx series, but I wonder if all the secrecy was worth it in terms of Eyefinity adoption by the gaming community (gamers, game makers, publishers, hardware folks) vs what it could have been with a full-blown announcement from the beginning. Obviously that would have led to a big rush from Nvidia to compete, maybe a delay of Fermi just to add 3 display outputs.
 
Man, reading articles by anand always makes me feel like I've gotten smarter. Even if I didnt. :D

The money quote, about Eyefinity:

"His desire to do this wasn’t born out of pure lunacy, Carrell does have a goal in mind. Within the next 6 years he wants to have a first generation holodeck operational. A first generation holodeck would be composed of a 180 degree hemispherical display with both positionally and phase accurate sound."


All the secrecy around Eyefinity makes it sound like they got the jump not only on Nvidia, but their own company and the game vendors as well. Calling it a different codename with each OEM to trace leaks, now thats planning. A bit early for a post-mortem on the 5xxx series, but I wonder if all the secrecy was worth it in terms of Eyefinity adoption by the gaming community (gamers, game makers, publishers, hardware folks) vs what it could have been with a full-blown announcement from the beginning. Obviously that would have led to a big rush from Nvidia to compete, maybe a delay of Fermi just to add 3 display outputs.

yeah I was blown away by that first statement. ATI has a plan for the furture and it will be glorious. The whole secrecy thing was well written and felt like some kind of white collar counter intelligence movie. you could replace ATI with CIA and sunspot with nuke/bioweapon plans it would be a pretty decent movie.
 
Awsome article. The section on "SunSpot" was very intresting.

I read that and the first thing that popped into my mind was "skin cancer lesions."

AFAIK Eyefinity still isn't fully implemented?
 
Back
Top