Overclock.net › Forums › Industry News › Hardware News › [BusinessKorea]Samsung and SK Hynix to Supply High Bandwidth Memory to Nvidia
New Posts  All Forums:Forum Nav:

[BusinessKorea]Samsung and SK Hynix to Supply High Bandwidth Memory to Nvidia

post #1 of 55
Thread Starter 
Quote:
According to industry sources on Sept. 20, Samsung Electronics and SK Hynix are planning to mass produce second-gen High Bandwidth Memory (HBM) for Nvidia's Pascal GPU starting in the first quarter of next year, after completing pilot production and reliability tests by the end of 2015. [

Earlier this year, SK Hynix was already supplying first-gen HBM to AMD and Nvidia ahead of Samsung. Micron uses Hybrid Memory Cubes (HMC), a technology similar to HBM, but falling behind it in productivity

An industry source said, "There are clear signs of a change in the structural design of CPUs, GPUs, DRAM memory, and storage led by Intel and Nvidia, which control standards in the market."
The source added, "Samsung and SK Hynix are both expanding their technical bases for various types of next-gen memory, including HBM chips

Source: http://www.businesskorea.co.kr/article/industry/manufacturing/12151/hynix-and-nvidia-samsung-sk-hynix-supply-high-bandwidth-memory

More news pointing out that Nvidia is getting ready with a release. Also I find it very interesting that SK Hynix have already supplied Nvidia with HBM1. Will that be used on an upcoming Pascal GPU which launch before the one with HBM2 later in 2016? thinking.gif
post #2 of 55
Besides capacity for double the VRAM and supposedly higher clock speeds, I can't see a difference between HBM1 and HBM2. I can understand the conservative approach by AMD when marketing HBM1. They are leaving open gains for marketing future releases. If you can control the heat, it is possible to get a two fold increase as it stands with LN2. The only setback is the current 4GB of Fiji. I predict AMD will release an 8GB variant before Pascal arrives. I also believe that future 8-hi stacks and higher capacities will have their trade offs in relation to speeds and latency.
Edited by anothergeek - 9/21/15 at 1:56pm
    
CPUMotherboardGraphicsGraphics
FX-8320 Gigabyte 990FXA UD3 Powercolor AX6850 Powercolor AX6850 
RAMHard DriveOptical DriveCooling
Corsair Vengeance LP CAS7 1600 (2 x 4GB) Samsung 840 Series Lite-On DVD-RW Corsair H80 
CoolingCoolingCoolingOS
Arctic Twin Turbo ll x2 Corsair AF140 x3 Corsair SP120 x2 Windows 7 Premium 64-bit 
MonitorKeyboardPowerCase
LG 32LH40 1080p 120Hz Interpolated 32" LCD Razer Arctosa Seasonic X650 Corsair 300R Windowed 
MouseMouse PadAudio
Razer Deathadder 2013 Razer Vespula Klipsch Promedia 2.1 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
FX-8320 Gigabyte 990FXA UD3 Powercolor AX6850 Powercolor AX6850 
RAMHard DriveOptical DriveCooling
Corsair Vengeance LP CAS7 1600 (2 x 4GB) Samsung 840 Series Lite-On DVD-RW Corsair H80 
CoolingCoolingCoolingOS
Arctic Twin Turbo ll x2 Corsair AF140 x3 Corsair SP120 x2 Windows 7 Premium 64-bit 
MonitorKeyboardPowerCase
LG 32LH40 1080p 120Hz Interpolated 32" LCD Razer Arctosa Seasonic X650 Corsair 300R Windowed 
MouseMouse PadAudio
Razer Deathadder 2013 Razer Vespula Klipsch Promedia 2.1 
  hide details  
Reply
post #3 of 55
Is getting ready? If they're planning mass production for Q1 you shouldn't expect Pascal with HBM2 until Q3/Q4.
post #4 of 55
Thread Starter 
Quote:
Originally Posted by Imouto View Post

Is getting ready? If they're planning mass production for Q1 you shouldn't expect Pascal with HBM2 until Q3/Q4.

Which is why I think it will be HBM1 which Nvidia apparantly have been buying for a while now.
Probably Pascal launch in Q1 with HBM1, then HBM2 cards with greater capacity in Q2/Q3 or something
post #5 of 55
Quote:
Originally Posted by iLeakStuff View Post

Which is why I think it will be HBM1 which Nvidia apparantly have been buying for a while now.
Probably Pascal launch in Q1 with HBM1, then HBM2 cards with greater capacity in Q2/Q3 or something

I would assume the HBM1 is for the mid and lower tier products that won't be getting more than 4bg to begin with. 4gb will probably be the new 2gb and any tier of cards that has less will probably be using GDDR5 with a refreshed gpu from the current lineup.
Computer
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7-4790k Gigabyte Z97X Gaming 7 Nvidia GTX 1080ti FE (EK block) CORSAIR Vengeance 
Hard DriveHard DriveCoolingOS
Samsung 850 Evo Mushkin Reactor Custom Loop Windows 10 
MonitorKeyboardPowerCase
Ben Q BL3201PT Logitech G110 EVGA 1300w Thermatake Core P5 
MouseAudio
LOGITECH G502 PROTEUS CORE Corsair 2100 
  hide details  
Reply
Computer
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7-4790k Gigabyte Z97X Gaming 7 Nvidia GTX 1080ti FE (EK block) CORSAIR Vengeance 
Hard DriveHard DriveCoolingOS
Samsung 850 Evo Mushkin Reactor Custom Loop Windows 10 
MonitorKeyboardPowerCase
Ben Q BL3201PT Logitech G110 EVGA 1300w Thermatake Core P5 
MouseAudio
LOGITECH G502 PROTEUS CORE Corsair 2100 
  hide details  
Reply
post #6 of 55
Quote:
Originally Posted by iLeakStuff View Post

More news pointing out that Nvidia is getting ready with a release. Also I find it very interesting that SK Hynix have already supplied Nvidia with HBM1. Will that be used on an upcoming Pascal GPU which launch before the one with HBM2 later in 2016? thinking.gif

nVidia really doesn't care about HBM1 much. HBM2 is what they are going to use for their government contracts and for their flagship designs. As you know it's twice as fast and has a ton more capacity. With HBM2 literally around the corner, why would nVidia bother to use HBM1 for anything other than maybe a low end card?

Anyone who has been following HBM really knew that AMD launching with 4GB of HBM1 really doesn't matter. It is just a stepping stone to HBM2 and 2016. nVidia is so far ahead, and their "outdated GTX 980Ti" still beats AMD's flagship with it, so why bother on this generation?

Once games start getting used to 6GB, 8GB, 12GB or GDDR5 and with 16GB+ of HBM2 ... even 4GB of HMB1 will be pretty much useless for anything but the lowest end graphics cards.

What the article speaks to me really is that it dispels the myth that AMD would be able to lock out nVidia with access to HBM1 and HBM2.
Edited by 47 Knucklehead - 9/21/15 at 3:39pm
post #7 of 55
Quote:
Originally Posted by 47 Knucklehead View Post

Quote:
Originally Posted by iLeakStuff View Post

More news pointing out that Nvidia is getting ready with a release. Also I find it very interesting that SK Hynix have already supplied Nvidia with HBM1. Will that be used on an upcoming Pascal GPU which launch before the one with HBM2 later in 2016? thinking.gif

nVidia really doesn't care about HBM1 much. HBM2 is what they are going to use for their government contracts and for their flagship designs. As you know it's twice as fast and has a ton more capacity. With HBM2 literally around the corner, why would nVidia bother to use HBM1 for anything other than maybe a low end card?

Throw some HBM1 on a 960/70/80 and shove it in a laptop.

I know that's what the 990M is supposed to be for, but seriously, you could solder the package direct to the motherboard like a BGA CPU and pack a heck of a lot of power into a tiny little laptop. 4GB should be plenty for that, and you could probably cut a 980 down to ~125w with the power savings.
Edited by KyadCK - 9/21/15 at 3:43pm
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
post #8 of 55
Quote:
Originally Posted by 47 Knucklehead View Post

What the article speaks to me really is that it dispels the myth that AMD would be able to lock out nVidia with access to HBM1 and HBM2.

That was a myth brought on by people here from reading 1 little article without understanding Samsung was already supplying Nvidia with HBM from long before the article surfaced. Mostly just all the people who didnt look at Nvidia's own slides they put out
Gaming
(17 items)
 
Gaming PC
(20 items)
 
 
CPUMotherboardGraphicsRAM
7700K AS Rock Z170 OC Formula Titan X Pascal 2050MHz 64GB DDR4-3200 14-14-14-34-1T 
Hard DriveHard DriveHard DriveCooling
950 EVO m.2 OS drive 850 EVO 1TB games drive Intel 730 series 500GB games drive Custom water cooling 
OSMonitorKeyboardPower
Win 10 Pro x64 AMH A399U E-Element mechanical, black switches, Vortex b... EVGA G3 1kw 
CaseMouseAudioAudio
Lian-Li PC-V1000L Redragon M901 LH Labs Pulse X Infinity DAC Custom built balanced tube amp with SS diamond ... 
Audio
MrSpeakers Alpha Prime 
  hide details  
Reply
Gaming
(17 items)
 
Gaming PC
(20 items)
 
 
CPUMotherboardGraphicsRAM
7700K AS Rock Z170 OC Formula Titan X Pascal 2050MHz 64GB DDR4-3200 14-14-14-34-1T 
Hard DriveHard DriveHard DriveCooling
950 EVO m.2 OS drive 850 EVO 1TB games drive Intel 730 series 500GB games drive Custom water cooling 
OSMonitorKeyboardPower
Win 10 Pro x64 AMH A399U E-Element mechanical, black switches, Vortex b... EVGA G3 1kw 
CaseMouseAudioAudio
Lian-Li PC-V1000L Redragon M901 LH Labs Pulse X Infinity DAC Custom built balanced tube amp with SS diamond ... 
Audio
MrSpeakers Alpha Prime 
  hide details  
Reply
post #9 of 55
Quote:
Originally Posted by KyadCK View Post

Throw some HBM1 on a 960/70/80 and shove it in a laptop.

I know that's what the 990M is supposed to be for, but seriously, you could solder the package direct to the motherboard like a BGA CPU and pack a heck of a lot of power into a tiny little laptop. 4GB should be plenty for that, and you could probably cut a 980 down to ~125w with the power savings.

Maybe. I think you'll see it (HBM1 and especially HBM2) more commonly used in laptops beyond just AMD and nVidia for graphics. Honestly, I think Intel would have more to gain with a lower power, cooler, and faster memory system for embedded and portable use.

Sure gaming laptops are nice, but most people don't use portables for that. Now having 16 or 32GB of HBM2 of memory in a laptop and have it accessed by both the video card and main memory ... everyone would really benefit from that.
post #10 of 55
Quote:
Originally Posted by EniGma1987 View Post

That was a myth brought on by people here from reading 1 little article without understanding Samsung was already supplying Nvidia with HBM from long before the article surfaced. Mostly just all the people who didnt look at Nvidia's own slides they put out

Yup, but the people who believe that refuse to listen to reason, that is why you have to beat them about the head and neck over and over again, with multiple sources, just to they realize just how foolish and out of touch with the industry they really are. Repetition is the only thing that really gets through to some of these people.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [BusinessKorea]Samsung and SK Hynix to Supply High Bandwidth Memory to Nvidia