Jump to content

Wait for Consumer Volta for Deep Learning at home or buy a Pascal?

So now that Volta has dropped, and we found out about the awesome uber powerful "Tensor Cores", is it better for someone learning Neural Networks, Machine Learning and Deep Learning to wait who knows how long until the Volta Consumer GTX's drop, or is it a too long wait and the Pascal (or even the upcoming Pascal Refresh GTX 20's ) are a better option? Can we expect the consumer GTX Voltas to have "Tensor Core" support? 

 

Also which of the following is a better option?

1. One GTX 1070/2070 in a few days, then add another GTX 2070 next year.

OR

2. Wait .... Wait..... Wait.... then buy a VOLTA GTX

 

P.S. Sorry for not suggesting 1080/2080's , if I buy one right now I'll be broke :S

 

Link to comment
Share on other sites

Link to post
Share on other sites

If you keep waiting you'll never own anything. Volta is so far off it shouldn't even be considered right now.

 

Having said that, Vega is weeks away and is worth waiting for at this point. The Vega Frontier edition is 13 TFLOPS of FP32 and 16GB of HBM2, that in-itself is worth seeing hit the market before making a major purchase decision.

 

 

Also, if Nvidia markets Volta anything like they did with Pascal you won't see all that fancy tech that's in the big deep-learning card in the consumer stuff. Just look at the P100 specs and compare it to what actually made it in the Xp.

LTT Unigine SUPERPOSITION scoreboardhttps://docs.google.com/spreadsheets/d/1jvq_--P35FbqY8Iv_jn3YZ_7iP1I_hR0_vk7DjKsZgI/edit#gid=0

Intel i7 8700k || ASUS Z370-I ITX || AMD Radeon VII || 16GB 4266mhz DDR4 || Silverstone 800W SFX-L || 512GB 950 PRO M.2 + 3.5TB of storage SSD's

SCHIIT Lyr 3 Multibit || HiFiMAN HE-1000 V2 || MrSpeakers Ether C

Link to comment
Share on other sites

Link to post
Share on other sites

I wouldn't expect consumer volta to have tensor cores. All that stuff takes up die space and doesn't help consumer workloads . they'll most likely be in the high end tesla cards ( gv100) , but  it'll be cut out from the consumer/lower-end workstation stuff ( gv102 , gv104 ) 

 

Same thing happened with FP64 ALU support in gp100 vs the rest of pascal .

 

Vega is supposed to have deep learning features , so that might be an option . Of course , you might need cuda support... 

AMD Ryzen R7 1700 (3.8ghz) w/ NH-D14, EVGA RTX 2080 XC (stock), 4*4GB DDR4 3000MT/s RAM, Gigabyte AB350-Gaming-3 MB, CX750M PSU, 1.5TB SDD + 7TB HDD, Phanteks enthoo pro case

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, nerdslayer1 said:

wait for vega 

pretty sure you need nvidia for deep learning. 

Main PC |CPU - i7-6700k|GPU - R9 290x tri-x 4gb|RAM - 16gb ddr4|MOBO - MSI z170 - A PRO|HDD - WD 1TB/240gb Sandisk |PSU - 700w Raidmax

Laptop |CPU - i7 4720hq|GPU - 960m 2gb|Ram - 8gb 2x4|Model - y50-70 Touch|SSD - 240gb Patriot drive|Display - 1920x1080 IPS touch

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, nerdslayer1 said:

 you need to finish school. 

woah my bad, its just based on what ive seen on basic research on it. 

 

for example i found this 

Graphics card/GPU

Which graphics card is the most important and the toughest question. For pretty much all machine learning applications, you want an NVIDIA card because only NVIDIA makes the essential CUDA framework and the CuDNN library that all of the machine learning frameworks, including TensorFlow, rely on.

Main PC |CPU - i7-6700k|GPU - R9 290x tri-x 4gb|RAM - 16gb ddr4|MOBO - MSI z170 - A PRO|HDD - WD 1TB/240gb Sandisk |PSU - 700w Raidmax

Laptop |CPU - i7 4720hq|GPU - 960m 2gb|Ram - 8gb 2x4|Model - y50-70 Touch|SSD - 240gb Patriot drive|Display - 1920x1080 IPS touch

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×