GeForce RTX 2080 and 2080 Ti - An Overview Thus far
Lian Li Strimer Review - RGB 24 Pin Extension Cable
Raijintek Orcus 240 AIO cooler review
AMD Ryzen Threadripper 2950X review
AMD Ryzen Threadripper 2990WX review
ASUS ROG SWIFT PG27UQ Monitor review
MSI B450i Gaming PLUS AC review
MSI Radeon RX 570 and 580 MECH 2 8G OC review
Preview: Unboxing Threadripper 2950X & 2990WX
Corsair SPEC Omega RGB review
AMD Working on GDDR6 DRAM Controller
It is always funny to see what people always note down on their linked in pages. This time it is a principal member of AMD's technical team, Daehyun Jun. His entry states he is/ was working on a DRAM controller for GDDR6 memory.
Currently, the linked-in page has been edited, and now just mentions "DRAM controller". It's all simply, HBM in any form and format is still expensive and difficult to use. If anything, the vega launch has shown exactly that. So it makes all the sense in the world for AMD to pursue a cheaper technology, and with GDDR6 being very promising, they started designing a new controller for their GPUs.
GDDR6 has the potential for double the bandwidth over current gen 8Gbps memory as used on say a GeForce GTX 1070 Ti. It's good news as we know it's coming likely on the next wave of graphics cards, also it uses less voltage and can theoretically double the bandwidth over current gen GDDR5 memory. To compare a little, current GDDR5 memory on say a Radeon RX 480 / GeForce GTX 1070 is roughly 7 or 8 Gbps, and GDDR5X on a GTX 1080 can do 10 or 11 Gbps.
GDDR6, when released, will start right where GDDR5X will end, at roughly 14 to maximum 16 Gbps. his means more bandwidth (up-to double over current bandwidth) and higher clock frequency memory. Voltages wise GDDR6 will be rated at 1.35v, which is exactly the same as GDDR5x. Other companies like SK Hynix and Micron are on track with GDDR6 as well.
GDDR6 will probably be available from early 2018.
AMD Working on Polaris 10 and Polaris 11 - 01/18/2016 10:26 AM
Though I am not sure about the significance of this one but AMD Radeon Technologies Group (RTG) head Raja Koduri slipped in an interview that they are not fabbing one, but two Polaris based GPUs....
AMD Working on 16-Core Processor with Integrated PCI Express 3.0 Controller - 01/17/2014 03:17 PM
In a document published by AMD some new information surfaced. It seems that AMD is working on a new 16-core processor, and that is a true 16 cores, thus not 8 cores and 4 GPU cores or anything. And no...
AMD working on new Ruby Technology Demo ? - 09/20/2013 11:28 AM
Back in the days AMD released the Ruby Technology demo whenever graphics cards where released. Who doesn't remember that ? Today an AMD press-release entered the good old mail-box showing a screensho...
AMD working on Radeon HD 3830 - 03/22/2008 11:04 AM
To be able to compete with NVIDIA's GeForce 8600 GTS in price and performance, AMD will launch a new model in the HD 3800 series with a more modest configuration than its two older counterparts. AMD'...
warlord
Senior Member
Posts: 1295
Joined: 2012-10-22
Senior Member
Posts: 1295
Joined: 2012-10-22
#5498152 Posted on: 12/05/2017 01:35 PM
Ram is important, but we need proper gpu juice to cope with. Resolution up to 8k is already smooth with that kind of bandwidth. I can't say the same about raw power of processing unit.
Some times progression of technology is not equal to all aspects.
Ram is important, but we need proper gpu juice to cope with. Resolution up to 8k is already smooth with that kind of bandwidth. I can't say the same about raw power of processing unit.
Some times progression of technology is not equal to all aspects.
Denial
Senior Member
Posts: 11440
Joined: 2004-05-16
Senior Member
Posts: 11440
Joined: 2004-05-16
#5498172 Posted on: 12/05/2017 02:41 PM
I wrote this in the other GDDR6 thread - faster RAM can lead to faster GPU indirectly. Previously an AMD card may have required a 256bit bus to hit x bandwidth, but with GDDR6 they may only need 128bit bus to hit that same x. This leads to reduced power consumption, which then can be used for faster clocks and it reduces the die size taken up by the memory controller - which potentially means you can stuff more cores in the same size chip.
Ram is important, but we need proper gpu juice to cope with. Resolution up to 8k is already smooth with that kind of bandwidth. I can't say the same about raw power of processing unit. Some times progression of technology is not equal to all aspects.
Some times progression of technology is not equal to all aspects.I wrote this in the other GDDR6 thread - faster RAM can lead to faster GPU indirectly. Previously an AMD card may have required a 256bit bus to hit x bandwidth, but with GDDR6 they may only need 128bit bus to hit that same x. This leads to reduced power consumption, which then can be used for faster clocks and it reduces the die size taken up by the memory controller - which potentially means you can stuff more cores in the same size chip.
warlord
Senior Member
Posts: 1295
Joined: 2012-10-22
Senior Member
Posts: 1295
Joined: 2012-10-22
#5498188 Posted on: 12/05/2017 03:45 PM
I agree technically with how you place it, but think it as that mate:
We have a motorway or a big wide road -> translates into bus depth (bits)
The quality of this road, surface, asphalt, safety precautions etc -> translates into bandwidth and ram generation
But in the end of the day, the most game changing factor is the kind of car, the year of it (model) and most of all its horsepower and/or capabilities. (GPU core for this subject)
All in all, I believe we firstly should have the demand for a GPU core that surely needs that kind of support, rather than paying for dram manufacturers innovation and their new products for testing. The whole GPU's hardware should be scaling equally as the years passing by.
I wrote this in the other GDDR6 thread - faster RAM can lead to faster GPU indirectly. Previously an AMD card may have required a 256bit bus to hit x bandwidth, but with GDDR6 they may only need 128bit bus to hit that same x. This leads to reduced power consumption, which then can be used for faster clocks and it reduces the die size taken up by the memory controller - which potentially means you can stuff more cores in the same size chip.
I agree technically with how you place it, but think it as that mate:
We have a motorway or a big wide road -> translates into bus depth (bits)
The quality of this road, surface, asphalt, safety precautions etc -> translates into bandwidth and ram generation
But in the end of the day, the most game changing factor is the kind of car, the year of it (model) and most of all its horsepower and/or capabilities. (GPU core for this subject)
All in all, I believe we firstly should have the demand for a GPU core that surely needs that kind of support, rather than paying for dram manufacturers innovation and their new products for testing. The whole GPU's hardware should be scaling equally as the years passing by.
Silva
Senior Member
Posts: 682
Joined: 2013-06-04
Senior Member
Posts: 682
Joined: 2013-06-04
#5498190 Posted on: 12/05/2017 03:49 PM
Plus: smaller bus equals to cheaper cards.
I wrote this in the other GDDR6 thread - faster RAM can lead to faster GPU indirectly. Previously an AMD card may have required a 256bit bus to hit x bandwidth, but with GDDR6 they may only need 128bit bus to hit that same x. This leads to reduced power consumption, which then can be used for faster clocks and it reduces the die size taken up by the memory controller - which potentially means you can stuff more cores in the same size chip.
Plus: smaller bus equals to cheaper cards.
schmidtbag
Senior Member
Posts: 3093
Joined: 2012-11-10
Senior Member
Posts: 3093
Joined: 2012-11-10
#5498215 Posted on: 12/05/2017 05:15 PM
HBM is great, practical, and arguably necessary, but only for compute tasks. I recall people here saying that Vega doesn't take advantage of HBM and that is simply false, when doing certain GPGPU tasks. Pretty much every time a Vega 64 outperforms a 1080Ti, that's because HBM kicked in.
HBM is overkill and a needless expense for gaming purposes. I seriously hope both AMD and Nvidia commit to HBM for FirePro/Quadro/Titan GPUs, but for everyone else's sake, GDDR6 is obviously a better choice. Until HBM can be mass-produced affordably, I don't want to see it on consumer/gamer GPUs.
HBM is great, practical, and arguably necessary, but only for compute tasks. I recall people here saying that Vega doesn't take advantage of HBM and that is simply false, when doing certain GPGPU tasks. Pretty much every time a Vega 64 outperforms a 1080Ti, that's because HBM kicked in.
HBM is overkill and a needless expense for gaming purposes. I seriously hope both AMD and Nvidia commit to HBM for FirePro/Quadro/Titan GPUs, but for everyone else's sake, GDDR6 is obviously a better choice. Until HBM can be mass-produced affordably, I don't want to see it on consumer/gamer GPUs.
kruno
Senior Member
Posts: 246
Joined: 2017-09-25
Senior Member
Posts: 246
Joined: 2017-09-25
#5498266 Posted on: 12/05/2017 08:31 PM
HBM is great, practical, and arguably necessary, but only for compute tasks. I recall people here saying that Vega doesn't take advantage of HBM and that is simply false, when doing certain GPGPU tasks. Pretty much every time a Vega 64 outperforms a 1080Ti, that's because HBM kicked in.
HBM is overkill and a needless expense for gaming purposes. I seriously hope both AMD and Nvidia commit to HBM for FirePro/Quadro/Titan GPUs, but for everyone else's sake, GDDR6 is obviously a better choice. Until HBM can be mass-produced affordably, I don't want to see it on consumer/gamer GPUs.
Power efficiency is second big HBM gain, you simply can not go against physics . HBM is stacked and very close to GPU so consumes much less power then classical memory. I have seen estimates on internet that says that if AMD had gone GDDR5 route it would cost them 100W of power more in comparison to HBM2
HBM is great, practical, and arguably necessary, but only for compute tasks. I recall people here saying that Vega doesn't take advantage of HBM and that is simply false, when doing certain GPGPU tasks. Pretty much every time a Vega 64 outperforms a 1080Ti, that's because HBM kicked in.
HBM is overkill and a needless expense for gaming purposes. I seriously hope both AMD and Nvidia commit to HBM for FirePro/Quadro/Titan GPUs, but for everyone else's sake, GDDR6 is obviously a better choice. Until HBM can be mass-produced affordably, I don't want to see it on consumer/gamer GPUs.
Power efficiency is second big HBM gain, you simply can not go against physics . HBM is stacked and very close to GPU so consumes much less power then classical memory. I have seen estimates on internet that says that if AMD had gone GDDR5 route it would cost them 100W of power more in comparison to HBM2
icedman
Senior Member
Posts: 751
Joined: 2013-02-22
Senior Member
Posts: 751
Joined: 2013-02-22
#5498272 Posted on: 12/05/2017 08:44 PM
AMD's problem in my opinion is they decided to go full retard on the gpu core a 3096 shader gpu with higher clocks would have probably been more efficient and probably performed the same while keeping costs down.
AMD's problem in my opinion is they decided to go full retard on the gpu core a 3096 shader gpu with higher clocks would have probably been more efficient and probably performed the same while keeping costs down.
kruno
Senior Member
Posts: 246
Joined: 2017-09-25
Senior Member
Posts: 246
Joined: 2017-09-25
#5498287 Posted on: 12/05/2017 09:33 PM
From all the stuff and all the website that i visit they all come to same conclusion that AMD :
A) doesn't have resources to develop 2 (or more)different arch's, one purely for graphics and second for compute ,so they develop one jack of all trades that is not very efficient
B) Process node. GloFo 14nm that has been co-developed with Samsung (Samsung was in charge) is optimized for low power low clocks (phones).And that process is great with power if you stay with in it's limits, and if you go over bye bye game over, power draw hit the roof
C) what you have said, if they (AMD) would lover the clocks then VEGA would be great power wise (you have bunch of examples and revives on line where they tried to under clock and under volt Vega and results are fantastic)
AMD's problem in my opinion is they decided to go full retard on the gpu core a 3096 shader gpu with higher clocks would have probably been more efficient and probably performed the same while keeping costs down.
From all the stuff and all the website that i visit they all come to same conclusion that AMD :
A) doesn't have resources to develop 2 (or more)different arch's, one purely for graphics and second for compute ,so they develop one jack of all trades that is not very efficient
B) Process node. GloFo 14nm that has been co-developed with Samsung (Samsung was in charge) is optimized for low power low clocks (phones).And that process is great with power if you stay with in it's limits, and if you go over bye bye game over, power draw hit the roof
C) what you have said, if they (AMD) would lover the clocks then VEGA would be great power wise (you have bunch of examples and revives on line where they tried to under clock and under volt Vega and results are fantastic)
Reddoguk
Senior Member
Posts: 1500
Joined: 2010-05-26
Senior Member
Posts: 1500
Joined: 2010-05-26
#5498323 Posted on: 12/05/2017 11:46 PM
Whats also interesting on that page is DDR5 in 2018 but by the look of it it doesn't start to get any faster till 2020. Really DDR5 already? Will that mean new ram sockets will be needed or will current DDR4 Dimms still work with DDR5?
Whats also interesting on that page is DDR5 in 2018 but by the look of it it doesn't start to get any faster till 2020. Really DDR5 already? Will that mean new ram sockets will be needed or will current DDR4 Dimms still work with DDR5?
user1
Senior Member
Posts: 893
Joined: 2016-01-29
Senior Member
Posts: 893
Joined: 2016-01-29
#5498343 Posted on: 12/06/2017 02:01 AM
Known for quite some time, references to Gddr6 have been present in the adl libraries for almost 2 years now
I will say , the implementation of the infinity fabric on amd's gpus, should make future gpu development faster and cheaper, they can basically just cut and paste various blocks(like memory controllers) now.
Known for quite some time, references to Gddr6 have been present in the adl libraries for almost 2 years now
I will say , the implementation of the infinity fabric on amd's gpus, should make future gpu development faster and cheaper, they can basically just cut and paste various blocks(like memory controllers) now.
sykozis
Senior Member
Posts: 20507
Joined: 2008-07-14
Senior Member
Posts: 20507
Joined: 2008-07-14
#5498350 Posted on: 12/06/2017 02:47 AM
I agree technically with how you place it, but think it as that mate:
We have a motorway or a big wide road -> translates into bus depth (bits)
The quality of this road, surface, asphalt, safety precautions etc -> translates into bandwidth and ram generation
But in the end of the day, the most game changing factor is the kind of car, the year of it (model) and most of all its horsepower and/or capabilities. (GPU core for this subject)
All in all, I believe we firstly should have the demand for a GPU core that surely needs that kind of support, rather than paying for dram manufacturers innovation and their new products for testing. The whole GPU's hardware should be scaling equally as the years passing by.
It's better to have memory that is faster than what is required by the GPU, than to have a GPU that requires data faster than the memory can handle it. In other words, the GPU itself should be the sole bottleneck of a graphics card, never the memory. If the memory is the bottleneck, the engineers screwed up.
Expecting GPU makers and memory makers to develop products that perfectly compliment each other is insane and would increase prices to the point that the dedicated graphics market would almost completely collapse.
I agree technically with how you place it, but think it as that mate:
We have a motorway or a big wide road -> translates into bus depth (bits)
The quality of this road, surface, asphalt, safety precautions etc -> translates into bandwidth and ram generation
But in the end of the day, the most game changing factor is the kind of car, the year of it (model) and most of all its horsepower and/or capabilities. (GPU core for this subject)
All in all, I believe we firstly should have the demand for a GPU core that surely needs that kind of support, rather than paying for dram manufacturers innovation and their new products for testing. The whole GPU's hardware should be scaling equally as the years passing by.
It's better to have memory that is faster than what is required by the GPU, than to have a GPU that requires data faster than the memory can handle it. In other words, the GPU itself should be the sole bottleneck of a graphics card, never the memory. If the memory is the bottleneck, the engineers screwed up.
Expecting GPU makers and memory makers to develop products that perfectly compliment each other is insane and would increase prices to the point that the dedicated graphics market would almost completely collapse.
Neo Cyrus
Senior Member
Posts: 8848
Joined: 2006-02-14
Senior Member
Posts: 8848
Joined: 2006-02-14
#5498373 Posted on: 12/06/2017 05:45 AM
Gee, maybe they'll actually sell video cards now instead of just saying "we don't make enough profit on HBM cards" and selling nothing.
It's about production cost. They don't want to go back to GDDR5 for their planned flagships, and they don't want to pay for HBM. They gambled on being able to produce it at lower costs by now and failed. In turn we all got screwed for it, as if the GPU market wasn't abysmal enough.
Gee, maybe they'll actually sell video cards now instead of just saying "we don't make enough profit on HBM cards" and selling nothing.
Ram is important, but we need proper gpu juice to cope with. Resolution up to 8k is already smooth with that kind of bandwidth. I can't say the same about raw power of processing unit. Some times progression of technology is not equal to all aspects.
Some times progression of technology is not equal to all aspects.It's about production cost. They don't want to go back to GDDR5 for their planned flagships, and they don't want to pay for HBM. They gambled on being able to produce it at lower costs by now and failed. In turn we all got screwed for it, as if the GPU market wasn't abysmal enough.
Amx85
Senior Member
Posts: 307
Joined: 2016-09-28
Senior Member
Posts: 307
Joined: 2016-09-28
#5499123 Posted on: 12/08/2017 02:39 PM
The weird thing is that now AMD will use GDDR and Nvidia HBM2 :V so WTF!
The weird thing is that now AMD will use GDDR and Nvidia HBM2 :V so WTF!
Loophole35
Senior Member
Posts: 9002
Joined: 2011-09-21
Senior Member
Posts: 9002
Joined: 2011-09-21
#5499127 Posted on: 12/08/2017 02:53 PM
I remember all the "experts" on this and other forums scoffing at 256bit cards. With GDDR6 you could have 780Ti bandwidth on 128bit memory bus.
Plus: smaller bus equals to cheaper cards.
I remember all the "experts" on this and other forums scoffing at 256bit cards. With GDDR6 you could have 780Ti bandwidth on 128bit memory bus.
Click here to post a comment for this news story on the message forum.
Senior Member
Posts: 2244
Joined: 2014-01-21
This could be Big News!