• 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.

AMD VEGA confirmed for 2017 H1

Status
Not open for further replies.
Courts literally agreed with arguments that it was not 8 core. Just because something is argued doesn't make it true. Obviously.

The 'entire' industry did not agree with this. You are speaking purely of AMD's arguments, which are clearly going to be one-sided.

The rest, man, this is just.....old stuff really. I cant honestly believe these arguments are still around, but I forget where I am sometimes. A forum where GPU fanboyism is rampant and AMD fans in particular are super vocal. I'm sure you'll spin my avoidance of arguing specifics as a win, but I'm doing so because it's tiring and a complete waste of time when the reality of things is obvious to anyone who doesn't have an agenda.

One quick example is the idea that Bulldozer was competitive in any/all multi-threaded tasks. Not true. As we saw going forward, tons of applications/games saw Intel's CPU's pull ahead with the same or lesser amount of cores/threads.

Last post on the subject I promise. The courts threw the case out and absolutely did NOT agree it wasn't an 8 core. The case was "did they lie with marketing about an 8 core", if the courts agreed the case would have been WON, not thrown out for stupidity. It got thrown out because every single industry expert agreed it was an 8 core.

If you want to win this case, you call up an Intel expert, an Nvidia expert, an ARM expert and some chip designers and have them testify it's a 4 core... no one would, not a single competitor of AMD ever called it a 4 core.

Lastly, I didn't say Bulldozer was competitive or ahead in all multithreaded applications. SO your example not surprisingly took a stance NO ONE anywhere took, and 'disproved it' thus apparently proving your point.

The concept, which is simple, is that if it was a 4 core then against another quad core with HT, it would gain NOTHING in performance between competing in single or four threaded applications to something with 8 threads, because there are no more cores to be used thus there is no advantage.

If an FX8350 is behind by 50% in a 4 threaded application but behind by 15% in an 8 threaded application against another chip that has 4 cores with or without HT, the only logically explanation is that it has.... more real cores otherwise how did it gain performance?

If it's behind by 30% in 4 threads and ahead by 20% in 8, or behind by 80% in 4 threads and behind by 60% in 8 threads, it's all an indication that more actual physical cores are being utilised. Nothing anywhere shows that Bulldozer is not an 8 threaded core. There is no argument that it isn't.

YOu are flat out lying here, the idea that "you can't believe these arguments are still around" is insane. What happened was, it was an 8 core, the industry and everyone in the world, every review site all called (and still do) an 8 core. At one time some idiot decided to sue AMD because they said it was only a 4 core. THe industry laughed, review sites laughed and explained why it wasn't. The court case was thrown out, the court case got ZERO support by a single player who would like to see AMD get hurt. The very argument that Bulldozer was only a 4 core was tiny, short lived and laughed at by the entire freaking world.... but you think this means exactly the opposite, that everyone believes it's a 4 core and only a few people at one time argued it was an 8 core.

There is literally no way to be more wrong about this than you are being, you've got it 100% reversed.

E-mail anandtech, e-mail Intel, or Nvidia, and ask them if Bulldozer is a 4 or 8 core.

As someone else said to you, if a single thread can run through an integer core, it can't share a core with a second thread and the chip can run 8 threads, it's an 8 core chip.


EDIT:- This is what makes it hard for people, someone spouts some utter utter nonsense, if you don't reply then other people start to believe it, if you reply you get called out for going off topic. There is no winning here when people start making absurd claims because it gets ignored and ends up misleading lots of people. Like the constant Nvidia fanboy diatribe that AMD drivers suck, it gains so much traction it's simply a 'fact' at this stage and almost impossible to argue against to most people because they've simply read it everywhere for so long.
 
Last edited:
EDIT:- This is what makes it hard for people, someone spouts some utter utter nonsense, if you don't reply then other people start to believe it, if you reply you get called out for going off topic. There is no winning here when people start making absurd claims because it gets ignored and ends up misleading lots of people. Like the constant Nvidia fanboy diatribe that AMD drivers suck, it gains so much traction it's simply a 'fact' at this stage and almost impossible to argue against to most people because they've simply read it everywhere for so long.

Pretty much nail on head. the Intel damage control team is trying to downplay blender and handbrake as good benchmarks. When they fairly demonstrate the ipc improvements of Zen.

We even had a good number of Nvidia shills trying to blast AMD for a small Vega preview and using titan xp's when the event was about Ryzen.
 
Are you two ^^^^ taking the ####? no one will buy anything that is overpriced by about 50% so they ain't going to be £600.
 
Yep, won't be cheap with HBM2 on em.

Because of course Fury X was 30% more than a 980ti due to having HBM 1 memory on, which was produced in less fabs, with no competitive production, in a tiny fraction of the supply HBM2 will have.... oh, wait.

8GB of GDDR5x probably cost £40, HBM2 might cost £70.... it's not going to turn a £400 card into a £800 card. Nvidia on the other hand take a card that should be £300 and price it to 600, despite not having HBM2 on it.

Memory won't massively increase the cost of the card regardless of what you use, unless you decided to put 32GB on it for no reason.

Fury cost more than a 290x, because it was a much bigger core and much lower volume, it also had a much more expensive cooling solution.
 
Most of the initial Fury x price was due to a number of things;

1. Limited stock leading to price gouging due to,
2. Needing an entirely new production chain due to,
3. New tech in the form of HBM and the need for interposers.

Interposers have been used in the past, a few pentiums had them between CPU Die and substrate.
 
If top Vega beats the TXP and has 16GB of HBM2 then I'll gladly pay just shy of £1000 for one knowing that it'll last me a fair while with my 3840x1600 res :) :cool:

Best be bloody good though!!! :D :p
 
Most of the initial Fury x price was due to a number of things;

1. Limited stock leading to price gouging due to,
2. Needing an entirely new production chain due to,
3. New tech in the form of HBM and the need for interposers.

Interposers have been used in the past, a few pentiums had them between CPU Die and substrate.

And the all-in-one cooler.
 
Status
Not open for further replies.
Back
Top Bottom