• 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.
Sure, but in the future AMD should take precautions not to put every egg in the same basket; they could have added a GDDR5X backup version in case of longer than expected lead times and released those as the "low-end" SKUs to compete with the gtx 1070 and kept HBM2 for GTX1080+ performance and up.
This. Why not take precaution ? Is that hard thing to do for amd i wonder
 
Looks like we might have a partial picture of one of the RX Vega cards:

https://www.techpowerup.com/233478/amd-radeon-vega-frontier-edition-spotted-in-amds-labs

If you look at the backplate and the position it looks nothing like that of any of the Fury based single or dual GPU cards. The closest is the Fury Nano but that has a different top bit.

This. Why not take precaution ? Is that hard thing to do for amd i wonder

Maybe because AMD has diverted most of their R and D in the last two years towards Zen development - AMD needs to make both CPUs and GPUs,so I suspect they decided CPU R and D has taken priority so they could not have a backup plan if HBM2 hit production issues. In the end most of their losses have been down to CPU side of the business,so its not surprising.
 
I don't see the problem to be honest. It is a joke for god sakes. Bit sad getting upset about such things imo :p

Regarding HBCC, what I understood is, to get full benefit of it the developer can design the game to make use of it using larger datasets. It would be like an option in game menus that will be greyed out unless you have hbcc for example. Otherwise it still works on some older games anyway. Which is what was expected is it not?

No, he said that to unlock the full potential you need to use bigger datasets. But they saw improvements even for current applications, that don't yet use extra large datasets.
 
I wonder if the RX Vega line will look different to the FE? Certainly at the very least it will be a different colour. Red instead of blue maybe? :)

No, he said that to unlock the full potential you need to use bigger datasets. But they saw improvements even for current applications, that don't yet use extra large datasets.

Yeah, probably this. But I would not be surprised if a developer they partner with would make a game making use of the HBCC feature somehow to it's fullest extent.
 
So that would mean VEGA will be a 8gb card with 480 GB/s bandwidth? The 1080ti has 484GB/s with 11gb of RAM.

Maybe im being stupid, but what was the big advantage of HBM2 over standard memory? is the latency better or something? cos it doesnt sound like its a major leap over existing tech, but costs a lot more.

I guess VEGA will handle memory somewhat different, but in 4k we can already see 6-8gb being used.

The current HBM2 is not at full speed, full speed would be 512GB/s and that is only for two stacks, if you use 4 stacks then it doubles again.
 
No, he said that to unlock the full potential you need to use bigger datasets. But they saw improvements even for current applications, that don't yet use extra large datasets.
Pretty sure that was true for the scenario where they used a vram-limited GPU (should help bring the price down if they decide to release a 4Gb Vega which should theoretically perform as good as an 8Gb or 12Gb card). :)
 
They also say they saw increased minimums in other titles that weren't optimized; so it could still be very nice.

[]To realize the full potential of HBCC, yes we will need to see content from game developers use larger datasets. But we have seen some interesting gains even on current software, particularly in min frame rates. Part of the goal of launching Radeon Vega Frontier edition, is to help speed up that process.

We don't know if he means all, some or even what percentage of older games show an improvements but minimum framerates is something AMD fall behind in.
We often see games tested where minimums are as much as 40% lower than the minimums on Nvidia cards also being tested. So,
If it creates a blanket minimums improvement in the majority of games that's very good news. Good minimums are essential.
 
We don't know if he means all, some or even what percentage of older games show an improvements but minimum framerates is something AMD fall behind in.
We often see games tested where minimums are as much as 40% lower than the minimums on Nvidia cards also being tested. So,
If it creates a blanket minimums improvement in the majority of games that's very good news. Good minimums are essential.

Well that's certainly game dependent anyway. Even the GTX 1070 shows nearly a 50% difference between it's own mins and averages in some games which is appalling.

Any improvement in minimums is always welcome to me, as I'd always much rather have 48FPS min and 67avg, than 38mins vs 80avgs
 
...
it will be a similar situation to the titan vs the rest of the geforce range. Frontier will have more grunt but slightly worse pcb and components than the custom rx vega cards.

i am specuating that frontier will lack the voltage range that RX vega has and that's why raja is dropping hints. frontier is probably bios locked at a lower range than the RX vega whilst frontier has the lower cloking chip but more functionality enabled for better FP performance.

More grunt? It better not, For me that'll be a deal breaker, I went with the cut-down Fury Pro because after a 6 month wait I was tired of waiting and didn't want to spend the next couple of months waiting for the fixed Fury X's or doing a Loadsa and buying different brands just to send them all back and end up out of pocket and on the Fury Pro anyway.
If AMD don't release the fully unlocked RX Vega first that's gonna be very disappointing too.

IMO worst case scenario in 8 weeks time you will have a Vega GPU in your machine, but I am willing to bet it will be in 4-6 weeks time.

I can live with that.
 
Well that's certainly game dependent anyway. Even the GTX 1070 shows nearly a 50% difference between it's own mins and averages in some games which is appalling.

Any improvement in minimums is always welcome to me, as I'd always much rather have 48FPS min and 67avg, than 38mins vs 80avgs

Depends on what kind of minimums we are talking about; a single 0.1ms drop to 38 fps from 80 would cause a minor annoyance while a 5s drop would be more cause for concern. Minimums are also highly dependent on the rest of the setup, especially when playing MP titles like BF1.
 
IMO worst case scenario in 8 weeks time you will have a Vega GPU in your machine, but I am willing to bet it will be in 4-6 weeks time.

I can live with that
Well that's certainly game dependent anyway. Even the GTX 1070 shows nearly a 50% difference between it's own mins and averages in some games which is appalling.

Any improvement in minimums is always welcome to me, as I'd always much rather have 48FPS min and 67avg, than 38mins vs 80avgs

My monitors a 75hz Freesync model so I'm happy with 74 fps max and minimums as high as possible, If I could game with minimums staying around or above 50/55/60 that would be ideal.
 
Depends on what kind of minimums we are talking about; a single 0.1ms drop to 38 fps from 80 would cause a minor annoyance while a 5s drop would be more cause for concern. Minimums are also highly dependent on the rest of the setup, especially when playing MP titles like BF1.

Aye. Better mins for life!

WCCFT's comment section is a treasure trove of trolls, baits, cringe, and worse haha.

Saw this being posted as well. People there going mental between each other.
c56ffab8ffcd454c9da3c10e45774389.png
 
Navi, multiple gpu cores, seen as a single large GPU all sharing their HBM2 as High-Bandwidth Cache; and allowing for improved SoC of CPU+GPU.

Just like Ryzen is multiple CCX linked via Infinity Fabric, AMD's goal by Navi is the same for GPUs. Raja spoke about in in 2016 where their end goal is for Multi-GPU cores to be as common as multi-CPU.

Infinity Fabric runs at 30-50GB/s on Ryzen, but goes all the way up to 512GB/s on Naples/Epyc and GPUs.

https://youtu.be/4qJj1ViyyPY?t=4m4s

I'm struggling to get my head around this part since GPU's are already have 1000's of small cores already on the die I don't get what advantage of having multiple smaller die's on a single die would bring (although one possible disadvantage could be increase delays timings).
 
Pretty sure that was true for the scenario where they used a vram-limited GPU (should help bring the price down if they decide to release a 4Gb Vega which should theoretically perform as good as an 8Gb or 12Gb card). :)

His exacts words were: "To realize the full potential of HBCC, yes we will need to see content from game developers use larger datasets. But we have seen some interesting gains even on current software, particularly in min frame rates." English is not my native language but it's obvious: content from game developers use larger datasets -> realize the full potential of HBCC.
 
Status
Not open for further replies.
Back
Top Bottom