7 ms·

But is 8 GB of GDDR whatever worth almost $1000? If AMD can put 16 in at about $500, paying another $1000 for an extra 8 is IMHO outrageous. Nvidia is charging

But is 8 GB of GDDR whatever worth almost $1000? If AMD can put 16 in at about $500, paying another $1000 for an extra 8 is IMHO outrageous. Nvidia is charging that much because many gamers hold to the “Intel & Nvidia are the best” even if benchmarks say otherwise.

malkia 6y ago
It's worth. There are really serious cases where extra GPU RAM is worth every penny. Not games, but say game development (or similar).
- ST2084 6y ago
  Resolve and other creative apps
pkulak 6y ago
Nvidia is also using much more expensive (and power-hungry) ram.
- AnthonyMouse 6y ago
  Which makes it more expensive, but it only makes it better if the expense actually translates into performance.
  - snvzz 6y ago
    >if it translates into performance. Which it doesn't. Thanks to AMD having some new, large "Infinity Cache" feature which they adapted from their Zen CPU architecture.
nightski 6y ago
To be fair to Nvidia it's not the same. AMD card is GDDR6 and 3090 is GDDR6X (which may be causing Nvidia's supply issues).
- snvzz 6y ago
  Doesn't matter much, when AMD is getting dramatically better performance out of their GDDR6 with their new 128MB "infinity cache". At the end, performance is what matters.
atty 6y ago
I do machine learning work and research, and when I upgrade I will pay the Nvidia “premium” without hesitation for that extra RAM and CUDA. I really wish it wasn’t so one-sided, I’d love to have a choice. Edit: should clarify that I’d really love to get a quadro or one of their data center cards, which aren’t gimped in certain non-gaming workloads... but I’m not made of money :)
- shock 6y ago
  > I do machine learning work and research, and when I upgrade I will pay the Nvidia “premium” without hesitation Would you still do it without hesitation if the money was coming of your own pocket?
  - Voloskaya 6y ago
    (Speaking for me) Yes. It's not 500$ extra for 8GB more, it's 500$ extra for being able to do ML at all basically. AMD GPUs have near zero support in major ML frameworks. There are some things coming out with Rocm and other niche things, but most people in ML already have enough work dealing with model and framework problems that using experimental AMD support is probably a no go. Hell, if AMD had a card with 8GB ram more than nvidia, and for 500$ cheaper, I would still go with nvidia. Everyone wish AMD would step their game up w.r.t ML workloads but it's just not happening (yet), Nvidia has a complete monopoly there.
    - delusional 6y ago
      This is more believable to me than the RAM being the decision maker. As an ML amateur it looks like most of it is built upon proprietary nvidia tech, which makes makes AMD a nonstarter, even if they did have a technically stronger card.
    - reilly3000 6y ago
      Would PlaidML be a viable option for you? https://github.com/plaidml/plaidml https://github.com/plaidml/plaidml
meragrin_ 6y ago
The 6900 XT the direct competitor to the RTX 3090. It is not $1000 cheaper. The 6900 XT was announced as $999 and the RTX 3090 is priced at $1499. The RTX 3090 also has raytracing hardware, support for machine learning, and other things. We have yet to see how the 6900 XT behaves in less than ideal conditions and performs with raytracing on. In the end, there might be other things which justify the extra $500.
sagarm 6y ago
It's product segmentation. NVidia dominates the ML space, so they don't have to worry much about competitive pressures. Increasing the efficiency of ML engineers is definitely worth at least $1500 them and their employers, so why leave money on the table?
- snvzz 6y ago
  >Increasing the efficiency of ML engineers is definitely worth at least $1500 them and their employers, so why leave money on the table? This would make sense if these were Titan cards with ML drivers. Instead, they are not, only the regular drivers are available, and FP32/64 performance is artificially capped. These aren't cards for ML.
  - nl 6y ago
    As an ML engineer I (and many others I know) would love them. The ML driver and FP32/64 performance capping aren't really issues since in reality we rarely hit those limits.

malkia 6y ago

It's worth. There are really serious cases where extra GPU RAM is worth every penny. Not games, but say game development (or similar).

ST2084 6y ago
Resolve and other creative apps

pkulak 6y ago

Nvidia is also using much more expensive (and power-hungry) ram.

AnthonyMouse 6y ago
Which makes it more expensive, but it only makes it better if the expense actually translates into performance.
- snvzz 6y ago
  >if it translates into performance. Which it doesn't. Thanks to AMD having some new, large "Infinity Cache" feature which they adapted from their Zen CPU architecture.

nightski 6y ago

To be fair to Nvidia it's not the same. AMD card is GDDR6 and 3090 is GDDR6X (which may be causing Nvidia's supply issues).

snvzz 6y ago
Doesn't matter much, when AMD is getting dramatically better performance out of their GDDR6 with their new 128MB "infinity cache". At the end, performance is what matters.

atty 6y ago

I do machine learning work and research, and when I upgrade I will pay the Nvidia “premium” without hesitation for that extra RAM and CUDA. I really wish it wasn’t so one-sided, I’d love to have a choice. Edit: should clarify that I’d really love to get a quadro or one of their data center cards, which aren’t gimped in certain non-gaming workloads... but I’m not made of money :)

shock 6y ago
> I do machine learning work and research, and when I upgrade I will pay the Nvidia “premium” without hesitation Would you still do it without hesitation if the money was coming of your own pocket?
- Voloskaya 6y ago
  (Speaking for me) Yes. It's not 500$ extra for 8GB more, it's 500$ extra for being able to do ML at all basically. AMD GPUs have near zero support in major ML frameworks. There are some things coming out with Rocm and other niche things, but most people in ML already have enough work dealing with model and framework problems that using experimental AMD support is probably a no go. Hell, if AMD had a card with 8GB ram more than nvidia, and for 500$ cheaper, I would still go with nvidia. Everyone wish AMD would step their game up w.r.t ML workloads but it's just not happening (yet), Nvidia has a complete monopoly there.
  - delusional 6y ago
    This is more believable to me than the RAM being the decision maker. As an ML amateur it looks like most of it is built upon proprietary nvidia tech, which makes makes AMD a nonstarter, even if they did have a technically stronger card.
  - reilly3000 6y ago
    Would PlaidML be a viable option for you? https://github.com/plaidml/plaidml https://github.com/plaidml/plaidml

meragrin_ 6y ago

The 6900 XT the direct competitor to the RTX 3090. It is not $1000 cheaper. The 6900 XT was announced as $999 and the RTX 3090 is priced at $1499. The RTX 3090 also has raytracing hardware, support for machine learning, and other things. We have yet to see how the 6900 XT behaves in less than ideal conditions and performs with raytracing on. In the end, there might be other things which justify the extra $500.

sagarm 6y ago

It's product segmentation. NVidia dominates the ML space, so they don't have to worry much about competitive pressures. Increasing the efficiency of ML engineers is definitely worth at least $1500 them and their employers, so why leave money on the table?

snvzz 6y ago
>Increasing the efficiency of ML engineers is definitely worth at least $1500 them and their employers, so why leave money on the table? This would make sense if these were Titan cards with ML drivers. Instead, they are not, only the regular drivers are available, and FP32/64 performance is artificially capped. These aren't cards for ML.
- nl 6y ago
  As an ML engineer I (and many others I know) would love them. The ML driver and FP32/64 performance capping aren't really issues since in reality we rarely hit those limits.