From ATI to AMD back to ATI? A Journey in Futility @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,510
From ATI to AMD back to ATI? A Journey in Futility - Trouble is brewing for AMD, Radeon Technologies Group, and Lisa Su, CEO and President. While on the surface it may just seem as though AMD engineers have once again missed the mark when it comes to building a new high-end graphics technology, but there is a lot more to the story.
 
All because they didn't give you a Nano.




/s :p


Been expecting something like this tbh. AMD has been failing at delivering anything really great for a while now. The 2xx series was OK, but way too power hungry for the performance (Same as their CPU's) and the 3xx series was just a refresh joke. Fury was just the same nonsense again as the 2xx series.
 
This reminds me of the office politics that killed Intel's XScale (fast low-power ARM CPU) group back when I worked there.

If those guys at AMD think moving to Intel will be their path to glory, they have no clue about the shark tank they are jumping into.
 
Kyle, this is some seriously first-class reporting. I didn't know that this was how things were going down internally. Great job dude.
Thank you sir. This is all bittersweet to be sure. We need AMD in the mix for the consumer, but it looks as if things will be going a different way. The Intel deal is not baked, so we will have to see what turns out. While all this lays at the feet of Raja and Lisa currently, there is no doubt in my mind that this course of action was set upon by Rory Read. He did damage to the GPU group that I do not think could be undone.
 
HardOCP's love of Nvidia and Intel being touted by fanbois, eh? I believe it. Somewhat ironic to them is that, interestingly, the first time I met Kyle in person was at an Eyefinity event downtown that [H] had partnered up on with AMD. :)

As for the subject matter of the article. I can see it. It wouldn't be the first time that blind ambition, arrogance, and absolute yet unfounded belief in ones own superiority leads to a huge downfall.
 
LMAO, not sure what else to say, not hard for AMD to seed rumor mills, remember back to the HD4xxx series, I do, very same type stuff was said about 4850(4870) being the top level of performance they were to have, lo and behold 4870 was released and it was far from a flop.

I personally have wished they had stuck using the name Ati Radeon as it was a familiar name, AMD for CPU Ati for GPU it just worked, RTG just sounds odd to me, internally it makes sense as it is its own division, IMO Ati had "pedigree" that was thrown aside for "Unification" which was not needed as they both are still their own pieces of a large puzzle.

I have a feeling Polaris 11 is what everyone has been basing their information on(like a 4850 as example good performance, good price, good power/temps) Polaris 10 will be a "4890" being top of range performance and come as a "surprise"
AMD has every reason to see false rumors about performance levels etc before the actual launch, cause claiming actual raw spec is suicide for sales and such, it also would cause Nvidia to possibly rush to reduce pricing or jump on releasing an even more powerful core which if P10 is just not quite at THAT level would also cause a major problem for AMD.

No matter, we will see soon enough either way, keeping in mind P10 was NOT designed to compete with GTX1080(officially or otherwise) so many keep forgetting this part and of course are LOL, and LMAO AMD sux type comments, I very much doubt AMD would or can afford to release actual true information, and from what I have seen so far Lisa Su and her Team have been spot on with launches and stuff when they have officially stated anything since she had taken over the reigns.

Great or just capable, either way, Radeon for me(when I can afford it) not going to see Nvidia in my machine, trick me once!
 
Somewhat ironic to them is that, interestingly, the first time I met Kyle in person was at an Eyefinity event downtown that [H] had partnered up on with AMD. :)
Great times and some great shows we did showcasing what Eyefinity could do for the gamer.
456490_383755424985808_2090628760_o.jpg
 
While I do find it a little "I didn't get invited to the party" you have a very good point. And that is very sad. I remember when AMD and ATI merged and the idea of them putting something special together to make a really good platform was great sounding. It's a shame both groups have done nothing but down hill since then. I've often wondered how the Radeon group would do if let to run independent.

Guess we'll find out. Good read.
 
Damn, i had been expecting AMD's demise to be sooner than 2020, but i didn't expect it to be this way... I know, i know, they aren't officially dead yet, and i am actually a huge fan, but this is getting beyond bad if true :S
 
Can't say I'll blame Koduri. He probably believes RTG will be significantly more competitive if he had the resources they need, resources which AMD can no longer afford to provide, and they probably will never be able to.

This is an industry that takes a lot of money to be competitive in, there's no any other way around this. I don't think AMD will ever be able to fully recover, they do not have the resources to be competitive. If Koduri believes it's a sinking ship, he sure as hell isn't sticking around.

Anyway thanks for sharing those info. [H] is certainly one of the few sites I trust to write about something like this only when there are actual facts to support it.
 
This reminds me of the office politics that killed Intel's XScale (fast low-power ARM CPU) group back when I worked there.

If those guys at AMD think moving to Intel will be their path to glory, they have no clue about the shark tank they are jumping into.

Exactly...it's akin to an independent game dev running into the iron maiden known as EA.

"Sure, we'll help ya! Just sign here. Yes, yes, good. Thanks for your company and all your IP. Now fuck off...you're all fired!"
 
I liked the article Kyle ;) I'm sure that this would make some people very jealous on this scoop ;).
If you check some of Renduchintala doings he cancelled 2 chips from Intel and the 3rd one he was going forward with. From this paywall article there might be some other light as to why and maybe it ties in with what you are suggesting ...
 
Interesting article - we should see in a month or two how well informed Kyle sources in AMD are but all of that sounds very probable.

This is terrible. If RTG goes to Intel, Nvidia won't have any competition. It'll be a one horse race.

Considering we are entering new era of lower level APIs that need to be optimised for each architecture that might be not so bad.
 
In the simplest terms AMD has created a product that runs hotter and slower than its competition's new architecture by a potentially significant margin

Well, the 1080FE is surely running very hot and throtteling, so I can't imagine how hot Polaris will be :(.

Also, we've known for months that Polaris is targeting 390/390x performance (so it's no surprise that the 1080/1070 have significantly better performance), are you implying that they won't even achieve that? That would be a really horrible engineering fail from AMD.
 
Very interesting scoop, Kyle - well done. If it was from anyone else, I'd be laughing at the epic trollbait.

If this is true, then my loss leader prediction a month or so ago, will be probably smack on. That's assuming they have worked improvements into Vega.

As a more AMD leaning type, the only [H] 'bias' I see, is Nvidia sometimes gets away with things lightly, in the same vein which AMD doesn't, e.g. Nvidia Win 10 drivers, however, no so much a gaming issue and thus not really something that'd pop up in an [H] review scope.
AMD CPUs are irrelevant for me unless making a super budget build, until Zen anyway if it remotely lives up to hype. Everything prior to my 2600k, from K6 days, was AMD.

RTG going to Intel or similar, is probably inevitable. One day a majority of systems will be system on chip.. progress if we like it or not. Maybe we'll still have something like an M.2 style discreet gpu, if the market is big enough..
 
Why would anyone want to see Intel buying out AMD graphics division as a bad thing. Times have changed. Tech is changing as well. People do not upgrade their cpu' s as often as they used to. Its a great thing for intel to have a strong graphics division This may improve the lack of competition you speak of. This is just rumors Mr. Bennett. Rumors spread by the folk that has been let go. :)
Could it be that Kuduri wants to cut off the AMD name from Graphics division because, to the uninformed it carries the weight of the not so stellar CPU performance.
 
Last edited:
huh so an intel fanboy works at AMD, and wants to use Intel cpu's instead of AMD's? makes sense AMD cpu's are not competitive anymore.

I will say that publishing this sort of article isn't going to solve your differences with AMD, only cause a greater divide. I know you don't care, either do I :)

Not sure what AMD's game plan is, but we find out soon enough.
 
Interesting, if rather dfisconcerting times. Partnering up with development on IGP tech with Intel (which I assume is the objective of that partnership) could be pretty lucrative, but it only accelerates what has become an increasing concern for many - a monopoly in the high-end desktop graphics market. Having seen some of the impact that situation has already had on desktop CPUs, it's not something any gamer should be excited about, fanboy or otherwise. I'm no fan of the way nvidia does business in a lot of ways, and to a lesser extent the same is true of Intels. I switched from using predominantly AMD graphics solutions to the GTX970 18 months or so ago due to displayport problems with running 2160p off a 290X. Now it's come to upgrade, AMD barely received a passing glance, there's nothing they either currently or soon will have to offer for a user in my category. I'd still have no issue recommending their midrange parts to people, but if we extrapolate current events - I wonder how long even doing that will be justified...


As far as the reporting goes, [H] is far from the only-site to get struck off AMD's list and having seen some of its reasons for doing so, I fully support being blunt, if the analysis is objective. There's a difference between baselessly slamming an organisation and bluntly pointing out their shortcomings. Reacting to the latter with hostility is never going to be PR positive.
 
That's pretty bleak. I'm seriously hoping for a Polaris/Zen 1-2 punch, because frankly I really really miss all-out performance wars like we saw with P4/Athlon, Core2/Phenom and GeForce 4000/Radeon 9xxx. I'm wishing that the Samsung buyout rumors were true. With the kind of R&D funding Sammy could throw at RTG, they could probably work some serious magic in the next couple of generations. Plus GCN getting stuffed into SoC packages for high end smartphones/tablets would be a neat thing to see.
 
No matter, we will see soon enough either way, keeping in mind P10 was NOT designed to compete with GTX1080(officially or otherwise) so many keep forgetting this part and of course are LOL, and LMAO AMD sux type comments, I very much doubt AMD would or can afford to release actual true information, and from what I have seen so far Lisa Su and her Team have been spot on with launches and stuff when they have officially stated anything since she had taken over the reigns.


Spot on with launches???? FuryX was supposed to crush 980ti, fail, Radeon Pro Duo was not even made available to reviewers but PC PERSPECTIVE found it to be a stuttery mess. They were supposed to be in the black a year ago and STILL losing millions every quarter. ZEN is now 2017, Vega 2017, Polaris NO actual launch dates, and now a reveal in MACAU???
Kyle may well be right!
 
It's clear that AMD is not firing on all cylinders, interesting to get a look into why that is.

Well, the 1080FE is surely running very hot and throtteling, so I can't imagine how hot Polaris will be :(.

Did you actually watch that video? The guy who made it doesn't even understand how the Turbo Boost 3 dynamic overclocking feature is designed to work and based on that ignorance makes a lot of very stupid statements.
 
Real journalists (like Kyle) shouldn't worry about pissing off one company or another. Their loyalty should be to their readers, not to these hardware manufacturers. This is one reason I appreciate [H] reviews.

PR is seemingly trying to twist Kyle's arm into getting "on message" like so many other review sites. He's not caving, which gives him tons of credibility in my eyes.
 
And the only thing AMD needed to do to avoid this was to invite [H] to Macau, but Noooooooooooooooooo. :rolleyes::rolleyes:
 
No matter, we will see soon enough either way, keeping in mind P10 was NOT designed to compete with GTX1080(officially or otherwise) so many keep forgetting this part and of course are LOL, and LMAO AMD sux type comments, Polaris NO actual launch dates, and now a reveal in MACAU???
Kyle may well be right!

What Kyle said is true in the article the miss steps are covered by price when they can't reach the performance target. There are way to many AMD cards that prove this is pretty normal ....
 
And the only thing AMD needed to do to avoid this was to invite [H] to Macau, but Noooooooooooooooooo. :rolleyes::rolleyes:
Hehe, this story have been in production just a bit longer than AMD's invites. We had planned to publish this before we were bound by any agreed NDA/Embargos. That said, not having an NDA / Embargo to worry about makes this easier to publish for sure.
 
A great piece of reporting, but the end result is troubling. This space has been bereft of any real competition for too long already. If that is set to continue for another generation at least, NVidia will not be shy about milking their position to the detriment of enthusiasts.
 
Intel acquiring a real GPU company that then gets help from intels engineers and direct access to intels top of the line fabs would be fucking amazing. This would cause nvidia's asshole to seriously pucker hard.
That's not how things work at Intel. You think getting wafer starts in the newest logic-process fabs is easy at Intel? Only if you are the x86 CPU guys. Funding, mindshare of management, resources of all kinds: it's hard as hell to get those at Intel if you aren't the main product line.

When was the last time an Intel acquisition didn't eventually just die or get sold off?
 
Back
Top