Navi and "Next Gen"

  • 70 results
  • 1
  • 2
Avatar image for Pedro
#1 Posted by Pedro (35486 posts) -

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

Avatar image for ronvalencia
#2 Edited by ronvalencia (28203 posts) -

@Pedro said:

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

Not correct. 5700 XT AE has 10.14 TFLOPS with 235 watts. https://videocardz.com/newz/amd-announces-radeon-rx-5700-xt-50th-anniversary-edition

AMD maxed out 5700 XT AE like RX-590.

Avatar image for Pedro
#3 Posted by Pedro (35486 posts) -

@ronvalencia: Last I checked I specifically stated 5700 XT not XT AE. Learn to read before smashing your face on the keyboard to respond.

Avatar image for ronvalencia
#4 Posted by ronvalencia (28203 posts) -

@Pedro said:

@ronvalencia: Last I checked I specifically stated 5700 XT not XT AE. Learn to read before smashing your face on the keyboard to respond.

There are 3 SKUs for 5700 series. Learn AMD's product stack for 5700 series.

Avatar image for evil_loli
#5 Posted by Evil_Loli (174 posts) -

Whatever this next-gen Xbox Scarlet will pack, I need to see what it can do before I shell out my cash. As for Ray tracing, meh.

Avatar image for Pedro
#6 Edited by Pedro (35486 posts) -

@evil_loli: Ray tracing is more a developer benefit than a gamer.

@ronvalencia said:

There are 3 SKUs for 5700 series. Learn AMD's product stack for 5700 series.

Learn to read. :)

Avatar image for ronvalencia
#7 Posted by ronvalencia (28203 posts) -

@Pedro said:

@evil_loli: Ray tracing is more a developer benefit than a gamer.

@ronvalencia said:

There are 3 SKUs for 5700 series. Learn AMD's product stack for 5700 series.

Learn to read. :)

Your knowledge is obsolete.

Avatar image for Pedro
#8 Posted by Pedro (35486 posts) -

@ronvalencia: Like your reading and comprehension skills. :)

Avatar image for ronvalencia
#9 Edited by ronvalencia (28203 posts) -

@Pedro said:

@ronvalencia: Like your reading and comprehension skills. :)

Navi 5700 XT AE is still Navi 5700 XT which didn't capped out at 9.75 TFLOPS.

Navi 5700 XT AE is effectively a reference design for AIB Navi 5700 XT OC cards and there could be faster OC "5700 XT" with 2Ghz clock speed.

Avatar image for davillain-
#10 Posted by DaVillain- (37589 posts) -

@Pedro said:

@evil_loli: Ray tracing is more a developer benefit than a gamer.

Learn to read. :)

I will agree with Pedro does have a point and ray-tracing needs to start somewhere. I don't expect ray-tracing to take place right away when these next-gen console's launch, the technology is gonna take a long time for developers to work on consoles.

Avatar image for Pedro
#11 Posted by Pedro (35486 posts) -

@ronvalencia: 5700 XT is capped at 9.75 TFLOPS. Yikes!

@davillain- said:

I will agree with Pedro does have a point and ray-tracing needs to start somewhere. I don't expect ray-tracing to take place right away when these next-gen console's launch, the technology is gonna take a long time for developers to work on consoles.

It would be interesting to see what other things they can do with the tech outside of lighting. If it aids in general ray casting applications it would be great addition.

Avatar image for goldenelementxl
#12 Posted by GoldenElementXL (3352 posts) -

@ronvalencia said:

Your knowledge is obsolete.

This is the post of the year. Ron posts like an evil AI from an 80's sci-fi movie

Avatar image for fedor
#13 Posted by Fedor (5277 posts) -

@Pedro said:

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

Yes.

Avatar image for ronvalencia
#14 Edited by ronvalencia (28203 posts) -

@Pedro said:

@ronvalencia: 5700 XT is capped at 9.75 TFLOPS. Yikes!

@davillain- said:

I will agree with Pedro does have a point and ray-tracing needs to start somewhere. I don't expect ray-tracing to take place right away when these next-gen console's launch, the technology is gonna take a long time for developers to work on consoles.

It would be interesting to see what other things they can do with the tech outside of lighting. If it aids in general ray casting applications it would be great addition.

"5700 XT" NAVI 10 XT didn't capped at 9.75Mhz since AE still has the same "NAVI 10 XT".

AMD wants the extra $50 for itself.

Avatar image for Pedro
#15 Posted by Pedro (35486 posts) -

@ronvalencia: That struggle. I stated 5700 XT and you keep complaining and talking about 5700 XT AE. Yikes!

Avatar image for ronvalencia
#16 Posted by ronvalencia (28203 posts) -

@Pedro said:

@ronvalencia: That struggle. I stated 5700 XT and you keep complaining and talking about 5700 XT AE. Yikes!

Your argument is flawed when there's another "5700 XT" SKU with higher clocks.

Avatar image for fedor
#17 Posted by Fedor (5277 posts) -

@ronvalencia said:
@Pedro said:

@ronvalencia: Like your reading and comprehension skills. :)

Navi 5700 XT AE is still Navi 5700 XT which didn't capped out at 9.75 TFLOPS.

Navi 5700 XT AE is effectively a reference design for AIB Navi 5700 XT OC cards and there could be faster OC "5700 XT" with 2Ghz clock speed.

LMAO!!! No there won't.

Avatar image for Pedro
#18 Posted by Pedro (35486 posts) -

@ronvalencia: Lets say we take the 5700 XT AE, its consumes more power and more power is not an option for consoles. So, back to the 5700 XT which is stated in the OP.

I look forward to your complaint. :)

Avatar image for ronvalencia
#19 Edited by ronvalencia (28203 posts) -

@Pedro said:

@ronvalencia: Lets say we take the 5700 XT AE, its consumes more power and more power is not an option for consoles. So, back to the 5700 XT which is stated in the OP.

I look forward to your complaint. :)

225 watts to 235 watts. it's only 10 watts difference.

IF it's like Vega II's over-voltage AMD lazy craftsmanship, under-voltage would have lowered 5700 XT AE's power consumption.

Read https://www.reddit.com/r/Amd/comments/ao43xl/radeon_vii_insanely_overvolted_undervolting/

AMD fanboys disappointed with AMD's lazy craftsmanship i.e. giving NVIDIA a free run.

Vega II holding higher clocks with under-voltage LOL. WTF is AMD doing? AMD giving NVIDIA free goals.

nvidia GPUs get unstable when undervolted. They are already being shipped with optimal voltage unless you want to OC.

you could improve Vega 64's efficiency by 34% already (the same value as with Radeon VII):

https://www.computerbase.de/2017-09/radeon-rx-vega-bios-power-test/2/#abschnitt_performance_pro_watt

The Radeon VII then only needs 207 watts while playing, whopping 70 watts less than in the factory state. Then the graphics card is also suddenly 22 watts more frugal than the GeForce RTX 2080 FE. While this can also reduce the voltage a bit, but since the Nvidia Maxwell generation works much closer to the optimum than AMD.

Avatar image for ellos
#20 Edited by ellos (2058 posts) -

Yes its all been about SSDs and going to stay that way. The marketing folks stayed away from TF talks. They cant sit there and explain why a similar TF navi is more powerful then same current TF measurement. They condition the market to hear Tera Flops. Now shit changed and the leap is no longer big. Sony will have to send out Cerny and MS tried to cover themselves by drooping lines like 4x more powerful then X. You gotta love marketing.

Avatar image for Pedro
#21 Posted by Pedro (35486 posts) -

@ellos said:

Yes its all been about SSDs and going to stay that way. The marketing folks stayed away from TF talks. They cant sit there and explain why a similar TF navi is more powerful then same current TF measurement. Sony will have to send out Cerny and MS tried to cover themselves by drooping lines like 4x more powerful then X. You gotta love marketing.

This is very true. They are probably struggling with the marketing because TFLOPS are going be less 2x the mid-gen refresh.

Avatar image for ronvalencia
#22 Posted by ronvalencia (28203 posts) -

@ellos said:

Yes its all been about SSDs and going to stay that way. The marketing folks stayed away from TF talks. They cant sit there and explain why a similar TF navi is more powerful then same current TF measurement. They condition the market to hear Tera Flops. Now shit changed and the leap is no longer big. Sony will have to send out Cerny and MS tried to cover themselves by drooping lines like 4x more powerful then X. You gotta love marketing.

Well, TMU unit is twice the width in NAVI. Nothing is free.

Avatar image for locus-solus
#23 Posted by locus-solus (1124 posts) -

Navi 10 can't do ray tracing but Navi 20 is rumored to be able to. maybe the consoles will be based off Navi 20?

Avatar image for Zero_epyon
#24 Posted by Zero_epyon (13332 posts) -
@Pedro said:

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

Comparing TFlops between architectures is nearly useless since TFlops are an architecture specific measurement. Which is probably why MS and Sony aren't mentioning it because to the layman, it would appear that the power of the consoles is dropping next gen, when the opposite will be the case.

Next gen is supposed to be about 4K/60. You still need a GPU that has the muscle to push that out and currently, even the X doesn't deliver 4K/30 consistently.

Avatar image for Random_Matt
#25 Posted by Random_Matt (4294 posts) -

TF does not indicate performance. These buzzwords and numbers are merely for the clueless buyers. Next gen will be a downclocked XT in my opinion, 200W+ is most of the console already.

Avatar image for daredevils2k
#26 Posted by Daredevils2k (2107 posts) -

@Pedro said:

@ronvalencia: Like your reading and comprehension skills. :)

Avatar image for XVision84
#27 Posted by XVision84 (16332 posts) -

I'm glad there will be greater advancements with the SSD and CPU side. It means next gen won't just be better textures and prettier effects.

Avatar image for boxrekt
#28 Posted by BoxRekt (2004 posts) -

As others have said the TF measurement using old Radion measurement isn't equivalent to that real world performance of Navi.

According to DF with the teraflop calcualtions with Navi, you can add at least 25% to that measurement.

Loading Video...
Avatar image for Grey_Eyed_Elf
#29 Edited by Grey_Eyed_Elf (6486 posts) -

@boxrekt said:

As others have said the TF measurement using old Radion measurement isn't equivalent to that real world performance of Navi.

According to DF with the teraflop calcualtions with Navi, you can add at least 25% to that measurement.

Loading Video...

Which they did in their explanation.

The 5700XT is the full fat 40 CU chip for Navi... So by dissabaling 4 CU's for yield purposes would bring them to 36 CU's, on par with the 5700 Pro in terms of CU count but using XT's core clocks would put it at 8.4 TFLOPS and a 8.4 TFLOPS on Navi relates to 10.7 TFLOPS on Vega based on AMD's IPC improvements of 25% which would land in Vega 57/ GTX 1080 level of performance... But that depends on the core clocks it could be slower than 8.4 TFLOPS.

We won't see more than 40 CU's and we won't see a console with more than 9.7 TFLOPS using Navi since the TPD would be too much for them to lower. Even at 8.4 TFLOPS with Navi would require serious optimisation to get it at confortable TDP levels.

Avatar image for ronvalencia
#30 Posted by ronvalencia (28203 posts) -

@daredevils2k said:
@Pedro said:

@ronvalencia: Like your reading and comprehension skills. :)

I didn't do this.

Avatar image for ronvalencia
#31 Edited by ronvalencia (28203 posts) -

@Grey_Eyed_Elf said:
@boxrekt said:

As others have said the TF measurement using old Radion measurement isn't equivalent to that real world performance of Navi.

According to DF with the teraflop calcualtions with Navi, you can add at least 25% to that measurement.

Which they did in their explanation.

The 5700XT is the full fat 40 CU chip for Navi... So by dissabaling 4 CU's for yield purposes would bring them to 36 CU's, on par with the 5700 Pro in terms of CU count but using XT's core clocks would put it at 8.4 TFLOPS and a 8.4 TFLOPS on Navi relates to 10.7 TFLOPS on Vega based on AMD's IPC improvements of 25% which would land in Vega 57/ GTX 1080 level of performance... But that depends on the core clocks it could be slower than 8.4 TFLOPS.

We won't see more than 40 CU's and we won't see a console with more than 9.7 TFLOPS using Navi since the TPD would be too much for them to lower. Even at 8.4 TFLOPS with Navi would require serious optimisation to get it at confortable TDP levels.

Based on historical results during 14/16 nm era AMD GPUs, both MS and Sony paid for larger GPUs when compared to RX-480/RX-580 36 CU.

MS's X1X vs RX-580 basic parameters

  • X1X GPU is 93 percent from RX-580's base clock speed.
  • X1X GPU is 85 percent from RX-580's boost clock speed.
  • X1X GPU has 6 TFLOPS FP32 with lower power consumption when compared to RX-580's 6 TFLOPS 185 watts.
  • X1X GPU has 12 GDDR5-6800 chips to be powered vs RX-580's eight GDDR5-8000 chips. X1X has the superiority at 4K over RX-580.
  • The entire X1X box has about 185 watts.
  • MS paid for 44 CU size GPU which is 22 percent higher CU count when compared to RX-480/RX-580 36 CU.
  • Microsoft used a smart VRM design that auto-under-voltage and optimize the power delivery for the GPU. X1X's VRM design is not AMD's VRM design.

Sony's PS4 Pro vs RX-480 basic parameters

  • PS4Pro GPU is 81 percent from RX-480's base clock speed at 1120 Mhz
  • PS4Pro GPU is 72 percent from RX-480's boost clock speed at 1266 MHz
  • PS4Pro GPU has 4.2 TFLOPS FP32 with lower power consumption when compared to RX-480's 5.8 TFLOPS with ~150 watts.
  • Sony paid for 40 CU size GPU which is 11 percent higher CU count when compared to RX-480/RX-580 36 CU.
  • PS4 Pro's result is like AMD's gimped mobile Polaris RX 480 without boost mode.

MS beaten Sony on rivalling the basic TFLOPS power from PC desktop's RX-580 with X1X GPU consuming less power and higher 12 GDDR5 chip count when compared to AMD's RX-580 6 TFLOPS 185 watts configuration.

During 16 nm era AMD GPUs: MS has shown to have superior electrical engineering skills (smart VRM design) over AMD's and Sony's electrical engineering skills.

Vega II's dumb VRM design shows AMD's lazy VRM craftsmanship as shown from this example https://www.reddit.com/r/Amd/comments/ao43xl/radeon_vii_insanely_overvolted_undervolting/

If Vega II has X1X's smart VRM design, this end user "retail beta testing" problem/disappointment would be less.

If Sony learns from X1X, Sony should NOT use AMD's dumb VRM designs.

AMD sets the voltage on their cards higher than ideal to ensure that all the Vega cards can run stable at stock clocks. The bottom % of chips will require these high voltages to stay at stock clocks, so to get more chips out the door they set the voltages that high.

Most Vega cards do not need such high voltage, and dropping the voltage the efficiency increases due to lower heat, which allows the clock to boost higher and stay cooler.

Heat increases with voltage, thus lower voltage means lower temps, which also means higher clocks, which means better performance and efficiency.

---------

Xbox Scorpio has VRM design with automatic under-voltage feature profiling per GPU silicon.

Both MS and Sony could throw away the bottom % of chips that requires high voltages.

Both Microsoft and Sony shouldn't pretend their game consoles to be DTR 17 inch laptop parts in a game console box, but build proper desktop systems with higher TDP limit.

Avatar image for scatteh316
#32 Edited by scatteh316 (10252 posts) -

@Pedro said:

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

I think DF said it best in their latest video, Tflops can't be compared between the current and next generation consoles as the architectures are so different.

1 Navi Tflop is likely a decent chunk faster then 1 Tflop from PS4's now 7 year old Southern Islands based GPU.

So a Southern Islands based GPU would likely need 12Tflop+ to match and compete with a 9.75Tflop Navi based GPU.

The RT 'hardware' acceleration is a grey area, doing it via GPU hardware pixel shaders is technically hardware accelerating it but it's not hardware accelerated in the same sense RTX is.

So it's either a play on words and it is done via pixel shading or it is actually hardware accelerated like RTX is, only time will tell.

Avatar image for pc_rocks
#33 Posted by PC_Rocks (2559 posts) -
@ellos said:

Yes its all been about SSDs and going to stay that way. The marketing folks stayed away from TF talks. They cant sit there and explain why a similar TF navi is more powerful then same current TF measurement. They condition the market to hear Tera Flops. Now shit changed and the leap is no longer big. Sony will have to send out Cerny and MS tried to cover themselves by drooping lines like 4x more powerful then X. You gotta love marketing.

Indeed! Poor poor cows(yup, I mentioned cows not all consolites deliberately), they used to says 14-20TF for next gen then adjusted that to 10+TF in the last 6 months to now DC'ing 8TF is okay.

Quite funny because they just used to throw TF number against PC gamers with Nvidia.

Avatar image for tormentos
#34 Posted by tormentos (29236 posts) -

@ronvalencia said:

MS beaten Sony on rivalling the basic TFLOPS power from PC desktop's RX-580 with X1X GPU consuming less power and higher 12 GDDR5 chip count when compared to AMD's RX-580 6 TFLOPS 185 watts configuration.

MS didn't beat anyone dude stop your freaking ass kissing of MS they simply waiting 1 year more and snatched better hardware there is not rocket science behind it.

Loading Video...

The RX580 beating the xbox one X results.

1440p Ultra 60FPS for the most part the xbox one X has to drop all the way to 1080p to do 60FPS.

In 4k as well beat the xbox one X for most of the time,keep in mind FH4 is a game were the lead platform is the XBO X not PC,so you can clearly hit 1440p at basically 60FPS with a few drops,a considerable boost in resolution while having 4CU less and 4GB less of memory.

There is many games outthere were the RX580 beat the xbox one X even that it has 4CU less and less memory and bandiwidth by the way.

@Pedro said:

With Navi 5700 XT capping out at 9.75 TFLOPS with 40 CU (Xbox One X 40 and PS4 Pro 36) what are the chances of PS5 and Xbox Scarlet pushing beyond 9.75 TFLOPS and a greater number of CUs?

In addition to this, MS claims that will have hardware accelerated ray tracing. Navi, currently doesn't support dedicated hardware accelerated ray tracing but compute shading via shaders. Is MS claiming this method to be hardware accelerated? Are they going to have a custom solution for hardware ray tracing?

Is next gen going to be all about the SSD and CPU?

This consoles don't need to be held back by the last ones,one of the beneffits of 7nm is its dencity it can pack more CU while having lower TDP than previous GPU.

Alto i wonder what price this machine will have if it cross 10TF mark,the xbox one X still is $400+ and it has 6TF,alto as time move on you can get a stronger GPU for the same price as the xbox one X that exceed it in power.

I think the chances are high,but my concern is price and TDP as well.

Avatar image for Tessellation
#35 Posted by Tessellation (9021 posts) -

What i understand is that 40navi CUs for example are more powerful and faster than 40 CUs of the xbox one x.

Avatar image for rmpumper
#36 Edited by rmpumper (669 posts) -

@ronvalencia said:
@Pedro said:

@ronvalencia: That struggle. I stated 5700 XT and you keep complaining and talking about 5700 XT AE. Yikes!

Your argument is flawed when there's another "5700 XT" SKU with higher clocks.

How is that relevant when talking about consoles with downclocked components? Are you expecting the next gen consoles to have brick power supplies and use over 300W?

Avatar image for fedor
#37 Posted by Fedor (5277 posts) -

@Tessellation: Basically, yes. Navi is a far more refined architecture compared to Polaris.

Avatar image for osan0
#38 Posted by osan0 (15452 posts) -

hard to say.

the more i look at navi 10 the more i think "this is half of the RDNA architecture. Its RDNA shaders with a mixture of vega and polaris components for the more specialised parts (with some tweaks)". A lot of the material i see focuses on better shader performance and efficiency (which is critically important of course). but not a lot on tesselation (is that the geometry engine?), ROPs, TMUs and so on...the more specialised stuff in a GPU.

if the PS5 and X2 were coming out this year i would be a bit more concerned. but they (well the X2 anyway) are not due until Q4 2020. By then navi 20 should also be out which will tell us more.

the ray tracing stuff....yeah hard to say. i mean technically navi10 can do ray tracing (as can any GPU) so it could be just MS and sony technically being correct but.....taking the whizz.

Navi 20 may have a much better capability to deal with real time ray tracing which could also be in the consoles. its only rumours now but its not unheard of for consoles to get some architectural improvements before they come to the PC (the 360 had an earlier version of unified shaders that were the basis of the HD 2000 series for example). sony and MS may also have their own solution to be bolted onto to navi in some way too (though that could get messy).

what i think we can definitely (well probably) take away from navi now is the shader architecture. if you want to see the make up of a compute unit for the PS5 and X2 for rasterisation...there it is basically. there may be some more minor tweaks and improvements between now and next year but by and large there it is.

after that it still really is speculation.

Avatar image for phbz
#39 Posted by phbz (4597 posts) -

I love you Ron! *shows boobs*

Avatar image for rdnav2
#40 Edited by RDNAv2 (35 posts) -

5700 XT is only 251mm2 die space.

Zen2 CPU is expected to take up 70mm2 of die space.

Combined that’s 320mm2

Add more CU’s and lower the clocks.

Die Space for these new consoles is expected to be between 350-400mm2 according to rumors

Avatar image for X_CAPCOM_X
#41 Posted by X_CAPCOM_X (8660 posts) -

@phbz said:

I love you Ron! *shows boobs*

lmao

Avatar image for ronvalencia
#42 Edited by ronvalencia (28203 posts) -

@rmpumper said:
@ronvalencia said:
@Pedro said:

@ronvalencia: That struggle. I stated 5700 XT and you keep complaining and talking about 5700 XT AE. Yikes!

Your argument is flawed when there's another "5700 XT" SKU with higher clocks.

How is that relevant when talking about consoles with downclocked components? Are you expecting the next gen consoles to have brick power supplies and use over 300W?

Based on recent history,

MS's X1X vs RX-580 basic parameters

  • X1X GPU's base clock speed is 93 percent from RX-580's base clock speed.
  • X1X GPU's base clock speed is 85 percent from RX-580's boost clock speed.
  • X1X GPU has base clock speed 6 TFLOPS FP32 with lower power consumption when compared to RX-580's boost mode 6 TFLOPS 185 watts.
  • X1X GPU has 12 GDDR5-6800 chips to be powered vs RX-580's eight GDDR5-8000 chips. X1X has the superiority at 4K over RX-580.
  • The entire X1X box has about 185 watts.

Sony's PS4 Pro vs RX-480 basic parameters

  • PS4Pro GPU's base clock speed is 81 percent from RX-480's base clock speed at 1120 Mhz
  • PS4Pro GPU's base clock speed is 72 percent from RX-480's boost clock speed at 1266 MHz
  • PS4Pro GPU has base clock speed4.2 TFLOPS FP32 with lower power consumption when compared to RX-480's boost mode 5.8 TFLOPS with ~150 watts.

MS has beaten Sony on rivalling the basic TFLOPS power from PC desktop's RX-580 while X1X GPU consumes less power when compared to AMD's RX-580 6 TFLOPS 185 watts configuration.

Facts for 16 nm era AMD GPUs: MS has shown to have superior electrical engineering skills (smart VRM design) over AMD's and Sony's electrical engineering skills.

Vega II's dumb VRM design shows AMD's lazy VRM craftsmanship as shown from this example https://www.reddit.com/r/Amd/comments/ao43xl/radeon_vii_insanely_overvolted_undervolting/

If Vega II has X1X's smart VRM design, "end user beta testing" complaints would be less.

MS has superior VRM design skills.

Avatar image for ronvalencia
#43 Edited by ronvalencia (28203 posts) -

@rdnav2 said:

5700 XT is only 251mm2 die space.

Zen2 CPU is expected to take up 70mm2 of die space.

Combined that’s 320mm2

Add more CU’s and lower the clocks.

Die Space for these new consoles is expected to be between 350-400mm2 according to rumors

AMD claims "RDNA 2" to be released before year 2021, hence somewhere in year 2020.

RDNA V2 has 7nnm+ node.

Microsoft's "hardware accelerated" ray-tracing confirmation has in-directly indicated Scarlet's 7nm+ process node.

RDNA V2 = Next Gen RDNA

https://www.techspot.com/news/80237-tsmc-7nm-production-improves-performance-10.html

N7+ has identical yield rates to N7 and will steadily improve, while also offering a 20% increase to transistor density. There’s also a 10% performance uplift or 15% power efficiency increase. AMD will take advantage of the former in their fourth-gen Ryzen which they’ve confirmed to use TSMC’s 7nm+

7nm+ has 20 percent increase transistor density, 10% performance uplift or 15% power efficiency increase.

AMD practically maxed out 1st gen 7nm with RX 5700 XT AE at 1980 Mhz with 235 watts has forced AMD head towards the next 7nm+ node.

Avatar image for HalcyonScarlet
#44 Posted by HalcyonScarlet (8444 posts) -

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

Loading Video...

Avatar image for ronvalencia
#45 Edited by ronvalencia (28203 posts) -

@HalcyonScarlet said:

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

https://www.reddit.com/r/Amd/comments/c1xkd0/navi_10121421_and_lite_variants_spotted_by_apisak/

"NAVI 21" has been spotted with AMD's Linux drivers and it's coming.

RX 5700 series aka NAVI 10 replaces Polaris 10/20/30.

Avatar image for HalcyonScarlet
#46 Posted by HalcyonScarlet (8444 posts) -

@ronvalencia said:
@HalcyonScarlet said:

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

https://www.reddit.com/r/Amd/comments/c1xkd0/navi_10121421_and_lite_variants_spotted_by_apisak/

"NAVI 21" has been spotted with AMD's Linux drivers and it's coming.

RX 5700 series aka NAVI 10 replaces Polaris 10/20/30.

Considering how fast your response was, you didn't watch the video. This is coming from AMD themselves.

Avatar image for ronvalencia
#47 Edited by ronvalencia (28203 posts) -

@HalcyonScarlet said:
@ronvalencia said:
@HalcyonScarlet said:

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

https://www.reddit.com/r/Amd/comments/c1xkd0/navi_10121421_and_lite_variants_spotted_by_apisak/

"NAVI 21" has been spotted with AMD's Linux drivers and it's coming.

RX 5700 series aka NAVI 10 replaces Polaris 10/20/30.

Considering how fast your response was, you didn't watch the video. This is coming from AMD themselves.

It's AMD's marketing speak when AMD is late to the market. AMD's damage control video is not new.

AMD's price argument is weak when cheaper NVIDIA Super arrives in July 2019.

AMD refuse to acknowledge NAVI's wave32 being similar to CUDA's wave32.

Avatar image for HalcyonScarlet
#48 Posted by HalcyonScarlet (8444 posts) -

@ronvalencia said:
@HalcyonScarlet said:
@ronvalencia said:
@HalcyonScarlet said:

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

https://www.reddit.com/r/Amd/comments/c1xkd0/navi_10121421_and_lite_variants_spotted_by_apisak/

"NAVI 21" has been spotted with AMD's Linux drivers and it's coming.

RX 5700 series aka NAVI 10 replaces Polaris 10/20/30.

Considering how fast your response was, you didn't watch the video. This is coming from AMD themselves.

It's AMD's marketing speak when AMD is late to the market. AMD's damage control video is not new.

AMD's price argument is weak when cheaper NVIDIA Super arrives in July 2019.

AMD refuse to acknowledge NAVI's wave32 being similar to CUDA's wave32.

Ahh ha sure. Except they're not wrong. What is the point in implementing a feature in its infancy just for the sake of it.

Avatar image for ronvalencia
#49 Edited by ronvalencia (28203 posts) -

@HalcyonScarlet said:
@ronvalencia said:

It's AMD's marketing speak when AMD is late to the market. AMD's damage control video is not new.

AMD's price argument is weak when cheaper NVIDIA Super arrives in July 2019.

AMD refuse to acknowledge NAVI's wave32 being similar to CUDA's wave32.

Ahh ha sure. Except they're not wrong. What is the point in implementing a feature in its infancy just for the sake of it.

AMD's aim for hardware accelerated ray-tracing is lighting effects. AMD is not even aiming for multiple different ray-tracing types e.g. lighting and reflections. AMD's ray-tracing argument is geared towards a single ray-tracing type and do it well.

RTX Turing can easily handle single type ray-tracing.

Ray-tracing also needs high memory bandwidth and higher VRAM storage which is gimped on RTX 2060 6GB. RTX 2060 Super has 8GB 256 bit GDDR6-14000 memory.

Ray-tracing should be used with variable shading rate to conserve shader resource and reduce memory bandwidth usage.

Avatar image for pc_rocks
#50 Posted by PC_Rocks (2559 posts) -

@HalcyonScarlet said:

Ray Tracing in that sense was left out on purpose with Navi. After RTX, AMD decided that more work was needed especially for that range. They were listening to gamers on the PC, the idea of Navi is high performance at a lower price.

Loading Video...

LOL sounds like DC from AMD. Their performance is no where near Nvidia even though they are on 7nm already and hence can't afford to do ray-tracing/DLSS in hardware. If they could they would have done it in a heart beat besides Navi was created to compete against 1080 and have been on the road map way way longer than RTX was a thing. So, it's just PR that they listened to the gamers and CHOSE not to implement RT.