[WCCF] NVIDIA Ampere GA100 GPU Rumored Specifications Detailed – 8192 CUDA Cores, Up To 48 GB HBM2e Memory, Up To 2.2 GHz Clocks & 300W TDP - Page 2 - Overclock.net - An Overclocking Community
Forum Jump: 

[WCCF] NVIDIA Ampere GA100 GPU Rumored Specifications Detailed – 8192 CUDA Cores, Up To 48 GB HBM2e Memory, Up To 2.2 GHz Clocks & 300W TDP

Reply
 
Thread Tools
post #11 of 87 (permalink) Old 03-06-2020, 04:44 PM
Food Editor
 
skupples's Avatar
 
Join Date: Apr 2012
Location: Bradentucky
Posts: 24,511
Rep: 719 (Unique: 378)
finally gonna get a titan over $9000 series?>

spend more save more !

copper switches are flipping amazing, gonna go play more speed typing games. its that good.

Add me on Steam, same name
R.I.P. Zawarudo, may you OC angels' wings in heaven.
If something appears too good to be true, it probably is.
Best R0ach Quote of all time : TLDR: Haswell might be the last legit gaming platform unless mice get their own non-USB interface on some newer architecture.
skupples is offline  
Sponsored Links
Advertisement
 
post #12 of 87 (permalink) Old 03-06-2020, 06:03 PM
New to Overclock.net
 
guttheslayer's Avatar
 
Join Date: Apr 2015
Posts: 3,869
Rep: 111 (Unique: 65)
Quote: Originally Posted by ToTheSun! View Post
Choo-choo!

Seriously, though, this can't be true. If someone can do it, it's daddy Jensen. BUT... even for a "leak", that's WAY too out there.
TBH the leak is expected since we already saw a 124 CU variant leaking around, and the improved boost clock is pretty much expected from 12nm to 7nm+ (2 node jump)

That being said, I dont believe both will come together, especially at full flat 128 CU with 2.2 GHz boost, it could happen but a 30 TFLOP seem to be alot more realistic, maybe a top 124 CU variants running at 1.9 GHz boost? That itself is impressive since we seem a almost 100% perf per watt improvement in the high end range.


And dont get your hope too high up, the GA100 would not land on your hand unless you willing to spur $3000 for their Titan A.

guttheslayer is offline  
post #13 of 87 (permalink) Old 03-06-2020, 06:09 PM
sudo apt install sl
 
Join Date: Dec 2009
Posts: 6,499
Rep: 188 (Unique: 130)
Quote: Originally Posted by ToTheSun! View Post
Choo-choo!

Seriously, though, this can't be true. If someone can do it, it's daddy Jensen. BUT... even for a "leak", that's WAY too out there.
If the SLIDE algorithm for CPU AI training becomes common nVidia will need a monster like this to compete.

Silent
(20 items)
CPU
Core i9 9900K... CoffeeTime! @ 5.1Ghz w/ 1.36v
Motherboard
Maximus VIII Formula
GPU
Titan RTX @ 2100Mhz/2075Mhz
RAM
TeamGroup Xtreem 32GB 3200Mhz CL15
Hard Drive
HP EX950 2TB
Hard Drive
Samsung 850 Evo 1TB
Hard Drive
Samsung 850 Evo 1TB
Power Supply
EVGA SuperNova 1200w P2
Cooling
EK Supremacy Full Copper Clean
Cooling
XSPC D5 Photon v2
Cooling
Black Ice Gen 2 GTX360 x2
Cooling
EK-Vector Radeon VII - Copper + Plexi
Case
Thermaltake Core X5 Tempered Glass Edition
Operating System
Clear Linux
Monitor
Acer XF270HUA
Keyboard
Cherry MX Board 6.0
Mouse
Logitech G600
Mouse
Alugraphics GamerArt
Audio
Definitive Technology Incline
Audio
SMSL M8A
▲ hide details ▲
WannaBeOCer is online now  
Sponsored Links
Advertisement
 
post #14 of 87 (permalink) Old 03-06-2020, 06:27 PM
Food Editor
 
skupples's Avatar
 
Join Date: Apr 2012
Location: Bradentucky
Posts: 24,511
Rep: 719 (Unique: 378)
Quote: Originally Posted by guttheslayer View Post
TBH the leak is expected since we already saw a 124 CU variant leaking around, and the improved boost clock is pretty much expected from 12nm to 7nm+ (2 node jump)

That being said, I dont believe both will come together, especially at full flat 128 CU with 2.2 GHz boost, it could happen but a 30 TFLOP seem to be alot more realistic, maybe a top 124 CU variants running at 1.9 GHz boost? That itself is impressive since we seem a almost 100% perf per watt improvement in the high end range.


And dont get your hope too high up, the GA100 would not land on your hand unless you willing to spur $3000 for their Titan A.
i'd buy titans again, if the gap were bigger than two core clusters, and faster memory.... but since I only game with em' nope nope nope! it would need to be a major difference.

Add me on Steam, same name
R.I.P. Zawarudo, may you OC angels' wings in heaven.
If something appears too good to be true, it probably is.
Best R0ach Quote of all time : TLDR: Haswell might be the last legit gaming platform unless mice get their own non-USB interface on some newer architecture.
skupples is offline  
post #15 of 87 (permalink) Old 03-06-2020, 06:53 PM
Invalid Media
 
J7SC's Avatar
 
Join Date: Dec 2018
Posts: 1,613
Rep: 73 (Unique: 31)
Re. SLIDE AI / DL, I read that article with great interest at Tom's HW a couple of days ago. But it's really early days and no apparent commercial adoption (yet). Besides, there maybe room for both matrix multiplications (GPU) and hashing (SLIDE / CPU), depending on the specific data AI / DL tasks.

Be that as it may, I put a NVLink 2080 Ti 8k Superposition run in the spoiler...of course NVLink is not the same as single card etc etc etc and in any case, this was just a test run. The point is about potential efficiency... dual 2080 Ti = 8704 CUDA cores and at 2.145 GHz, those sucked back a combined 760w (peak but fairly sustained) just for the two GPUs. If the 7nm 8192 CUDA GA100 can do this at 300w (or say even 360w with 120% PL), that would be an incredible step forward.

Finally, I don't think it is coincidence that this more detailed NVidia Ampere GA100 leak came out just when AMD had its Financial Analyst Day and presented RDNA 2 Yet with some serious impacts of covid-19 on the tech supply chains, who knows when we we actually get either AMD's RDNA 2 or NVidia's Titan or Ti versions of Ampere into our waiting (begging ?) hands.

Spoiler!
J7SC is offline  
post #16 of 87 (permalink) Old 03-06-2020, 06:56 PM
New to Overclock.net
 
dantoddd's Avatar
 
Join Date: Sep 2009
Posts: 1,578
Rep: 37 (Unique: 32)
too good to be true.

CPU
(4 items)
CPU
i7 8700k
GPU
RTX 2080Ti
RAM
16GB DDR4
Operating System
Windows 10 Pro
▲ hide details ▲
dantoddd is offline  
post #17 of 87 (permalink) Old 03-06-2020, 07:05 PM
sudo apt install sl
 
Join Date: Dec 2009
Posts: 6,499
Rep: 188 (Unique: 130)
Quote: Originally Posted by J7SC View Post
Re. SLIDE AI / DL, I read that article with great interest at Tom's HW a couple of days ago. But it's really early days and no apparent commercial adoption (yet). Besides, there maybe room for both matrix multiplications (GPU) and hashing (SLIDE / CPU), depending on the specific data AI / DL tasks.

Be that as it may, I put a NVLink 2080 Ti 8k Superposition run in the spoiler...of course NVLink is not the same as single card etc etc etc and in any case, this was just a test run. The point is about potential efficiency... dual 2080 Ti = 8704 CUDA cores and at 2.145 GHz, those sucked back a combined 760w (peak but fairly sustained) just for the two GPUs. If the 7nm 8192 CUDA GA100 can do this at 300w (or say even 360w with 120% PL), that would be an incredible step forward.

Finally, I don't think it is coincidence that this more detailed NVidia Ampere GA100 leak came out just when AMD had its Financial Analyst Day and presented RDNA 2 Yet with some serious impacts of covid-19 on the tech supply chains, who knows when we we actually get either AMD's RDNA 2 or NVidia's Titan or Ti versions of Ampere into our waiting (begging ?) hands.
Don't forget the researchers didn't utilize Intel's AVX SIMD instructions. As time goes on both Intel and AMD will keep improving their AI focused instructions.

Quote:
Optimizations

The researchers say that there are further performance improvements left as they have “just scratched the surface”. To that end, they say that they have not used vectorization – such as AVX SIMD instructions – including Intel’s DLBoost acceleration and claimed “there are a lot of other tricks we could still use to make this even faster.”
We know nVidia's release schedule already.

https://www.nvidia.com/en-us/deep-le...cts/titan-rtx/

GTC = New Tesla - https://www.nvidia.com/en-us/gtc/
Gamecom = Geforce
NIPS = Titan

Here's what a stock Titan RTX can do when running Resnet-50 benchmark when it's running at P0 with all of it's glorious features. I love having 24GB of memory!

Spoiler!

Silent
(20 items)
CPU
Core i9 9900K... CoffeeTime! @ 5.1Ghz w/ 1.36v
Motherboard
Maximus VIII Formula
GPU
Titan RTX @ 2100Mhz/2075Mhz
RAM
TeamGroup Xtreem 32GB 3200Mhz CL15
Hard Drive
HP EX950 2TB
Hard Drive
Samsung 850 Evo 1TB
Hard Drive
Samsung 850 Evo 1TB
Power Supply
EVGA SuperNova 1200w P2
Cooling
EK Supremacy Full Copper Clean
Cooling
XSPC D5 Photon v2
Cooling
Black Ice Gen 2 GTX360 x2
Cooling
EK-Vector Radeon VII - Copper + Plexi
Case
Thermaltake Core X5 Tempered Glass Edition
Operating System
Clear Linux
Monitor
Acer XF270HUA
Keyboard
Cherry MX Board 6.0
Mouse
Logitech G600
Mouse
Alugraphics GamerArt
Audio
Definitive Technology Incline
Audio
SMSL M8A
▲ hide details ▲
WannaBeOCer is online now  
post #18 of 87 (permalink) Old 03-06-2020, 07:08 PM
New to Overclock.net
 
Schmuckley's Avatar
 
Join Date: Jul 2011
Location: FL
Posts: 15,000
Rep: 709 (Unique: 566)
Schmuckley is offline  
post #19 of 87 (permalink) Old 03-06-2020, 07:20 PM
Invalid Media
 
J7SC's Avatar
 
Join Date: Dec 2018
Posts: 1,613
Rep: 73 (Unique: 31)
Quote: Originally Posted by WannaBeOCer View Post
(...) We know nVidia's release schedule already (...)

...fyi, I referenced the Titan Ampere version, not current Titan RTX re. dates. In any case, EY Consultancy just circulated a draft on the impact of covid-19 on the global tech industry to some major news outlets...apart from various conferences and roll-out events either being cancelled or shifted to online only, the actual physical manufacturing side is hugely affected....CPU/GPU, screens, RAM, M.2/SSD/HDD etc etc. I just don't want to see product roll-outs which are just paper tigers (which already happened plenty in '19) and no actual stock to purchase.
J7SC is offline  
post #20 of 87 (permalink) Old 03-06-2020, 07:24 PM
Overclocker
 
JackCY's Avatar
 
Join Date: Jun 2014
Posts: 10,820
Rep: 358 (Unique: 255)
Send the test samples over here
<==

Till then, a big rumor with no proof.
JackCY is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off