• 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.
N19h7m4r3 said:
Hold your horses folks. Videocardz.com looked into it. He has a list of all the 3DMark runs with that gpu I'D. Wccft only took the top one.


https://videocardz.com/70777/amd-radeon-rx-vega-3dmark11-performance#disqus_thread
Click to expand...
Interesting, Thought 1ghz mark for HBM2 was the max, that shows 1890 :confused:

Although that listing the stupid figure for gddr5 of 10000+ MHz. Always gets me that. I wish the mem would clock that high :p mind you same with sdram so....

oh look i've got a 62400 MHz CPU :eek:
Ok not comparable but you see what am getting at.
 
pete910 said:
Interesting, Thought 1ghz mark for HBM2 was the max, that shows 1890 :confused:

Although that listing the stupid figure for gddr5 of 10000+ MHz. Always gets me that. I wish the mem would clock that high :p mind you same with sdram so....

oh look i've got a 62400 MHz CPU :eek:
Ok not comparable but you see what am getting at.
Click to expand...
We think the memory is clocked at 945 MHz and 1890 just happens to be 2x that.
 
Schnippzle said:
Interesting to see that only two sites seem to be reporting "news".
Click to expand...

Wccft makes up news, fakes results, and talks pure rumors most of the time; along with stealing most of their images and information from Videocardz.

Videocardz is more responsible with leaks, and information; and only post things he's entirely sure of is legitimate.
 
N19h7m4r3 said:
Wccft makes up news, fakes results, and talks pure rumors most of the time; along with stealing most of their images and information from Videocardz.

Videocardz is more responsible with leaks, and information; and only post things he's entirely sure of is legitimate.
Click to expand...

While I agree that Videocardz are more reliable, I wouldn't say everything they post is 100% reliable.
 
Sargatanas2511 said:
While I agree that Videocardz are more reliable, I wouldn't say everything they post is 100% reliable.
Click to expand...

It's why I said is more responsible with leaks. Like just now with the RX card. He listed all the scores reported. Some are only overclocked 1070 levels; while when the card is overclocked it's just under overclocked AIB 1080.

wccft just took the took score and reported as if it's the "reference" speed at 1630Mhz.
 
Angilion said:
What evidence do you have that Vega RX will use a different and better GPU than Vega FE?
What evidence do you have that Vega RX will use different and better memory than Vega FE?
What evidence do you have that Vega RX will have more memory than Vega FE?
What evidence do you have that Vega RX will not be much cheaper than Vega FE?
What evidence do you have that Vega FE drivers are so seriously defective despite at least a year of development that huge gains are possible from drivers alone in a month?

If you think what I said is based on assumptions that are wrong, give a counter-argument.
Click to expand...

Equally ask what evidence do you have for your assumptions to show that they are correct or sufficient? Its the selection of information and assumptions to hold as true which is the issue.

Vega RX has at best the same GPU as Vega FE. Cant rule in or out a revised chip.
Vega RX has at best the same type of memory as Vega FE. Can't rule in or out memory being clocked a little higher (for likely minimal impact tbf).
Vega RX has at best the same amount of memory as Vega FE. Confirmed that stock config is with 8GB. No reason to assume an equal memory configuration to the FE 16GB as an optimal or best outcome.
Vega FE has what are stated to be gaming-specific drivers. I have running-specific shoes. I'm not setting any impressive times even if so-and-so runner could to do so in them.

It's not a huge gamble to use the FE performance as being indicative of RX performance because it's fundamentally the same card. If RX differs in hardware from FE, then RX will be worse than FE because it has to be much cheaper.
Assumption, other potential considerations - Retail cost is relevant but not as suggested. Enabling/disabling functionality and niche features doesn't have to make the product worse in unrelated use-cases. Certain features allow higher pricing when targeting specific markets. Useful for higher margins which help ROI and amortise the cost of r&d, no need to rely on gaming market alone to bear the responsibility. Every additional unit sold is increased economies of scale and extra revenue and profit (even if gpu is sold far cheaper than other variants and the margin is small).

It might previously have been theoretically possible that RX would use GPU and memory clocked significantly higher than FE, but the power requirement and heat generation of FE cards completely rules that out.
Assumption, other potential considerations: Fewer active dram dies = slightly less power; binning of silicon for better clockeing chips vs FE; larger custom HSF/AIO can handle more heat; maybe revised silicon; maybe features or operating parameters not optimal.

So any increase in performance would be down to an extra month's development on the drivers that have already been in developement for at least a year. It's...optimistic...to think that will make a huge difference to performance.
Assumption: In truth we know bugger all on driver status beyond it being the latest driver package approved for public release. Confirmed by Ryan Shrout from PCPerspective.

It comes down to this - if AMD does have something much better than Vega FE, why did they release Vega FE as the halo version at >£1000 a card?
That is assuming it is a Halo product. You are also thinking as though the gaming market is all that matters. We do know they set a deadline so would look to satisfy shareholders and close off a potential avenue for litigation. Releasing this niche model would enable heftier margins, even on poorer dies as all out performance not a priority (stability and features of greater priority and potential interest than peak performance gaming or otherwise).
Click to expand...

Incidentally I am not making any claims about the eventual performance of Rx Vega. All I know is that 'flavours' will be faster than the FE.
 
JediFragger said:
I'm wondering if they're not enabling due to the fear of getting sued the **** out of by nVidia!!
Click to expand...

Nvidia already had a lawsuit against Qualcom/Samsung about this. As far as I can tell the trial was going badly for Nvidia and ended up in them pulling out of the case (Samsung dropped a counter-suit against them).

However, I think the important finding was this ITC ruling quoted from the article I linked:

Since then, things have not progressed well for NVIDIA, with the US International Trade Commission (ITC) ruling that Samsung’s GPUs don’t infringe on NVIDIA’s patents, while also ruling that NVIDIA’s GPUs did infringe on Samsung’s patents.
Click to expand...

Note that the ruling mentions:

Specifically, the judge ruled that none of the relevant Andreo, PowerVR, or Mali GPUs infringe on NVIDIA’s vertex processing patent (7209140) or their patent on multithreaded execution of programs (7038685).
Click to expand...

Adreno is the tech Qualcomm bought from AMD...
 
akarypid said:
Nvidia already had a lawsuit against Qualcom/Samsung about this. As far as I can tell the trial was going badly for Nvidia and ended up in them pulling out of the case (Samsung dropped a counter-suit against them).

However, I think the important finding was this ITC ruling quoted from the article I linked:
Note that the ruling mentions:
Adreno is the tech Qualcomm bought from AMD...
Click to expand...

IIRC Nvidia recently lost a big patents case, and several of those patents were invalidated.
 
im guessing this has been mentioned before and sorry if it has, but can anyone else see this being, 1080 levels at 1070 pricing- but then amazing at MINING even with its huge power draw and end up being priced like a Ti...?????

seeing these benchmarks, personally just hope it allows Freesync 1440p owners to run their monitors at a competitive rates to G-sync products :)
sure there are more then a few that will be happy to run 144hz monitors at 100 plus at 1440p which 580s cant do and Fury has Memory limits
 
triss said:
Some of the 480 hype train thought 480 would be 980TI level Vega trains got to have some high dreams :)
Click to expand...

To be fair to that old hype train, it came down to clockspeed/power consumption.

People saw the GTX10 series and assumed (reasonably) that Polaris could hit similar clockspeeds.

Polaris is about 10% higher IPC than Hawaii/Fiji, so if it could hit ~1700MHz it would be in the ballpark of stock 980Ti performance.

Turns out GCN has a fundamental arch issue hitting high clockspeeds.

Can only hope Navi is a true ditching of GCN, since a lot of the strength of the 7nmLP process from GloFo/IBM is the high clockspeed potential. So they'd be throwing away a lot of the node's potential if they put another iteration of GCN on it. I'd expect a clockspeed-happy arch like Pascal would hit AT LEAST 2.5 GHz on it.
 
Angilion said:
What evidence do you have that Vega RX will use a different and better GPU than Vega FE?
What evidence do you have that Vega RX will use different and better memory than Vega FE?
What evidence do you have that Vega RX will have more memory than Vega FE?
What evidence do you have that Vega RX will not be much cheaper than Vega FE?
What evidence do you have that Vega FE drivers are so seriously defective despite at least a year of development that huge gains are possible from drivers alone in a month?

If you think what I said is based on assumptions that are wrong, give a counter-argument.
Click to expand...

Your arguments are based on assumptions and so would mine. So what's the point? You just don't know and neither do I.
 
Am happy with 1080 performance and still willing to buy. The pricing needs to be right though I wont be fooled into over paying.

Be a nice upgrade over my 290 still
 
Well looks Good enough for me to fork out 450 for it if thats the real score :) Thats descent jump frommmy overclocked 980ti that i plan to sell for 200-220.
 
The fastest results are maybe with more recent drivers. Also 30 mhz isn't overclock compared to 1080 which is running on 2ghz core in comparison lol
 
Status
Not open for further replies.
Back
Top Bottom