• 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.

Poll: ** The AMD VEGA Thread **

On or off the hype train?

  • (off) Train has derailed

    Votes: 207 39.2%
  • (on) Overcrowding, standing room only

    Votes: 100 18.9%
  • (never ever got on) Chinese escalator

    Votes: 221 41.9%

  • Total voters
    528
Status
Not open for further replies.
That bandwidth really intrigues me as well, as the reduction from 4 to 2 chips is what someone else mentioned as well when I was asking why the bandwidth was so low.

I actually wonder if they originally thought they would have 4 chips, but decided to reduce to 2 due to costs or availability of HBM2, and then screwed themselves for performance? Overclocking the memory on Vega should help see if that may be the case.

Hynix had issues getting HBM2 released in a timely manner,and if AMD had predicted that clockspeeds would be much higher than they expected it might mean they are bandwidth limited.

Sure the Draw Steam Binning Rasterizer might help a bit,but Nvidia does have more experience using it,especially since I believe it was developed for Tegra primarily(to save power).

It probably hasn't helped, bu AMD coudln't go with a 4 stack design again. HBM just isn't ready for GPUs yet.

The 1080ti offers 480GB/sec as it is.

ATM,it does seem Nvidia is still ahead in more effectively using bandwidth in gaming situations.
 
Isn't that 480 quid :p?

I don't really want to pay that money for that performance now. I could have done that last year.
Thats the dilemma, anything I buy now will last less time than it would have done had I bought last year.

Plus I don't want to buy Nvidia really :p. I'm happy to spite myself, but I'm not going to pay amd the same money for the same performance either

I posted the Quick Silver link twice, this is the one i was recommending for £400 https://www.overclockers.co.uk/pali...ddr5-pci-express-graphics-card-gx-03n-pl.html

Just hold your nose and dive in, its not all that bad, If anyone tells you nVidia drivers are better they are lying or have no idea... they really are not, but still, its very fast, its cool, efficient, the hardware is nice.
 
So... This just shows Vega FE had no gaming optimisations and what Gamer Mode will do in the next driver release.

00020-c2edd3842b45cd49.jpg
 
That bandwidth really intrigues me as well, as the reduction from 4 to 2 chips is what someone else mentioned as well when I was asking why the bandwidth was so low.

I actually wonder if they originally thought they would have 4 chips, but decided to reduce to 2 due to costs or availability of HBM2, and then screwed themselves for performance? Overclocking the memory on Vega should help see if that may be the case.


It was always designed around 2 chips. the costs of a 4 stack HBM2 and the large interposer is prohibitive. Even for the FE and Instinct cards they have just 2 stacks.
 
I posted the Quick Silver link twice, this is the one i was recommending for £400 https://www.overclockers.co.uk/pali...ddr5-pci-express-graphics-card-gx-03n-pl.html

Just hold your nose and dive in, its not all that bad, If anyone tells you nVidia drivers are better they are lying or have no idea... they really are not, but still, its very fast, its cool, efficient, the hardware is nice.

I do want to buy into Free-Sync too.
I'll just keep the 290x and whinge about it.

I can't stomach being 400 quid when my brother bought a 1070 msi for 370 last year.
 
I’m about 99% sure this is the reason for the delay. RTG driver team are having to rewrite the driver for TBR and it’s taking longer than expected.

So not exactly sandbagging as such, just a case of the card needing a few more weeks to ripen.

Can I just add to the posts above -

WAIT FOR BENCHMARKS

(there are a lot of you speculating but the fact is, the driver guys have two more weeks to sort this out)


AMD have already released benchmaks, you really expext AMD to be sandbagging performance. And 2 weeks isn't going to change anything , drivers have been in development for 18 months.

Time to face the facts.
 
AMD have already released benchmaks, you really expext AMD to be sandbagging performance. And 2 weeks isn't going to change anything , drivers have been in development for 18 months.

Time to face the facts.
Indeed, 2 weeks more won't bring 10-30% extra performance no matter how much you want it - otherwise AMD would have held off...
 
If they had something to show off, they would be showing it. I don't normally have this attitude, I just think they don't have a competitive product.
 
Also from the AT article on the launch:
http://www.anandtech.com/show/11680...-vega-64-399-rx-vega-56-launching-in-august/3

Talking to AMD’s engineers, what especially surprised me is where the bulk of those transistors went; the single largest consumer of the additional 3.9B transistors was spent on designing the chip to clock much higher than Fiji. Vega 10 can reach 1.7GHz, whereas Fiji couldn’t do much more than 1.05GHz. Additional transistors are needed to add pipeline stages at various points or build in latency hiding mechanisms, as electrons can only move so far on a single clock cycle; this is something we’ve seen in NVIDIA’s Pascal, not to mention countless CPU designs. Still, what it means is that those 3.9B transistors are serving a very important performance purpose: allowing AMD to clock the card high enough to see significant performance gains over Fiji.

Speaking of Fiji, there’s been some question over whether the already shipping Vega FE cards had AMD’s Draw Steam Binning Rasterizer enabled, which is one of the Vega architecture’s new features. The short answer is that no, the DSBR is not enabled in Vega FE’s current drivers. Whereas we have been told to expect it with the RX Vega launch. AMD is being careful not to make too many promises here – the performance and power impact of the DSBR vary wildly with the software used – but it means that the RX Vega will have a bit more going on than the Vega FE at launch.

So nearly 4 billion transistors were used to make the design clock better.

By the way the reads, Vega being a Fiji shrink seems to be somewhat correct, just that they had to make the die bigger than half of Fiji in order to get better clocks, 14nm Fiji at around half size might have only allowed it to hit something like 1300mhz
 
Seems you just want to see what you want to see. Point is one could have had this performance ages ago for the same price, not to mention on a card that runs more efficient. Vega came very late and did not improve price for performance like it should have. End of the day people complained Nvidia is milking us so people waited; over a year later now waiting those people will be rewarded by getting milked by AMD. lol.

14 months ago AMD were shouting about how out of order the price of a 1080 was ans how you could buy 2x 480's for the price of one 1080.
Fastforward to today...... need i say more.

Well at least I haven't bought Prey or Wolf II yet so that's something.
I'll just have to see if Nvidia do enough to entice me to buy a faster gpu between now and the release.
I tried Dying Light with my 1060 last night and it was a mess, I'm not sure why though,
The fps seemed okay after reducing just about everything and the memory usage got no higher than just over 5gb's so I'm not sure why it was so bad,
I haven't played it for months but back then it ran as smooth as butter with my Fury pro.
 
Beginning to think that just a die shrink for fiji with more memory would have been the better option whilst they worked on the true replacement (Navi).
 
Status
Not open for further replies.
Back
Top Bottom