4 ms·

I find the render network's ambition to be very compelling, and think the decentralized approach to rendering is necessary. Could something similar be accomplis

I find the render network's ambition to be very compelling, and think the decentralized approach to rendering is necessary. Could something similar be accomplished with a centralized model? Theoretically yes definitely. Practically though, my stance is this can't happen soon with open standards. IMO the project aims to solve a couple fundamental problems: - Graphical computing power is prohibitively expensive - Graphical computing power is inherently scarce and inaccessible, while most graphics resources are in reality sitting idle. - Thermal/power constraints are incompatible with a mass-market form factor for VR/AR/XR. Looking past the shilling in this tweetstorm I think it's the best summary of the core concepts I've seen: https://twitter.com/arbvision/status/1460774216136024071 https://twitter.com/arbvision/status/1460774216136024071 [edit: appended "open standards" above]

smoldesu 5y ago
Rendering-as-a-service already exists, even in a fairly distributed form.
- JayStavis 5y ago
  Certainly -- rendering services are great. I just think a decentralized approach wins on cost, accessibility, compatibility, and quality of service in certain instances. And I think the margin it wins by enables a whole new class of products.
  - high_byte 5y ago
    Arguably (actually probably) decentralization overhead costs more than centralized farm. To clarify for everyone else (building on top of parent comment) - distributed does not necessarily mean decentralized.
    - JayStavis 5y ago
      > decentralization overhead costs more than centralized farm. Say more :) Definitely with you distributed doesn't mandate decentralized and the two are easy to conflate.
      - high_byte 5y ago
        What else should I say? I think existing (centralized) render farms just work well. Although there's the $RNDR token which seems more like hype to me than actual improvement and again, arguably (probably) worse.
  - danielmarkbruce 5y ago
    How does it win? Can you step through in a concrete fashion how this would work in one instance you refer to? Highly distributed render farms exist, and they work under a central control. How does a blockchain do anything to reduce cost, increase compatibility or quality of service v existing solutions?
    - JayStavis 5y ago
      For the record I wouldn't argue "blockchain" is the only possible means of decentralization/distribution for offloading graphics compute. With that said, an example would be a solo artist making CGI in C4D for a client/hobby (this is a small subset of potential use cases, but most relatable today). Typical workflow costs would look like: - Local Workstation ($4K w/ 2 high end nvda gpus) - Render 2000 frames at 4K resolution for final render (need a better figure here but let's call it $500) - Time spent on workflow rendering - 5 hours - Time spent on final rendering - 10 hours In a decentralized render situation: - Local computer ($500 - in theory no discrete graphics necessary) - Render 2000 frames at 4K resolution ($100) - Time spent on workflow rendering - 1 hour - Time spent on final rendering - 2 hours I'm not trying to handwave at the data here btw, I just have a day job and want to answer promptly. RNDR is far and away cheaper at the moment than local or render farm compute: https://twitter.com/rsquires/status/1466599124451487744 https://twitter.com/rsquires/status/1466599124451487744 Here are the main reasons why it's cheaper: - Most decentralized GPU operators are "running at a loss" -- people who own GPUs don't expect a return so any dollars earned while they're idle is good with them. - No "datacenter tax". Server GPUs are objectively more expensive than the network's GPU's which bias towards consumer. - In terms of time, rendering is one of the most parallelizable workloads. 2000 frames rendered sequentially on a cloud instance is just slower than 2000 frames rendered in parallel on 50 cloud instances. For QoS, because it's so much cheaper, redundancy is much more justifiable to add to the system. On compatibility, an SDK would allow people to build their own products that leverage the compute of the network. That's not something there is a high incentive for centralized render farms to offer. If you want to go deep I highly recommend recent podcast with CEO of OTOY and the two hosts of the Mograph podcast: https://www.youtube.com/watch?v=SRtOMF4JKd4 https://www.youtube.com/watch?v=SRtOMF4JKd4
f00zz 5y ago
From your link, I found this a bit odd: > As video games get more photo realistic (...) the demand for rendering is going to grow immensely. I don't see what a decentralized render farm has to do with video games, it's never going to render at real-time rates.
- JayStavis 5y ago
  Agreed that is one of the hairier challenges a concept like this faces. I do think that argument is worth unpacking though. Would a render node in your neighborhood have acceptable latency? IIRC they have demoed "real-time" NFT's based on an Unreal Engine node streaming frames at acceptable e2e latency. Personally I'd think a bit broader than frame-streaming for achieving "console in the room" levels of video game fidelity.
  - duskwuff 5y ago
    The primary issue wouldn't be latency. It'd be access to assets (which can easily be tens to hundreds of gigabytes for a modern game) and upstream bandwidth (which is often as little as 5 Mbps for residential users). These aren't problems that a blockchain can solve. In fact, I'll go a step further and say: distributed graphics rendering, or other computation, isn't a problem which is made any easier at all by the application of a blockchain. It's yet another case of cryptocurrency enthusiasts taking a problem that people seem interested in and claiming, without justification, that The Blockchain Solves This.

smoldesu 5y ago

Rendering-as-a-service already exists, even in a fairly distributed form.

JayStavis 5y ago
Certainly -- rendering services are great. I just think a decentralized approach wins on cost, accessibility, compatibility, and quality of service in certain instances. And I think the margin it wins by enables a whole new class of products.
- high_byte 5y ago
  Arguably (actually probably) decentralization overhead costs more than centralized farm. To clarify for everyone else (building on top of parent comment) - distributed does not necessarily mean decentralized.
  - JayStavis 5y ago
    > decentralization overhead costs more than centralized farm. Say more :) Definitely with you distributed doesn't mandate decentralized and the two are easy to conflate.
    - high_byte 5y ago
      What else should I say? I think existing (centralized) render farms just work well. Although there's the $RNDR token which seems more like hype to me than actual improvement and again, arguably (probably) worse.
- danielmarkbruce 5y ago
  How does it win? Can you step through in a concrete fashion how this would work in one instance you refer to? Highly distributed render farms exist, and they work under a central control. How does a blockchain do anything to reduce cost, increase compatibility or quality of service v existing solutions?
  - JayStavis 5y ago
    For the record I wouldn't argue "blockchain" is the only possible means of decentralization/distribution for offloading graphics compute. With that said, an example would be a solo artist making CGI in C4D for a client/hobby (this is a small subset of potential use cases, but most relatable today). Typical workflow costs would look like: - Local Workstation ($4K w/ 2 high end nvda gpus) - Render 2000 frames at 4K resolution for final render (need a better figure here but let's call it $500) - Time spent on workflow rendering - 5 hours - Time spent on final rendering - 10 hours In a decentralized render situation: - Local computer ($500 - in theory no discrete graphics necessary) - Render 2000 frames at 4K resolution ($100) - Time spent on workflow rendering - 1 hour - Time spent on final rendering - 2 hours I'm not trying to handwave at the data here btw, I just have a day job and want to answer promptly. RNDR is far and away cheaper at the moment than local or render farm compute: https://twitter.com/rsquires/status/1466599124451487744 https://twitter.com/rsquires/status/1466599124451487744 Here are the main reasons why it's cheaper: - Most decentralized GPU operators are "running at a loss" -- people who own GPUs don't expect a return so any dollars earned while they're idle is good with them. - No "datacenter tax". Server GPUs are objectively more expensive than the network's GPU's which bias towards consumer. - In terms of time, rendering is one of the most parallelizable workloads. 2000 frames rendered sequentially on a cloud instance is just slower than 2000 frames rendered in parallel on 50 cloud instances. For QoS, because it's so much cheaper, redundancy is much more justifiable to add to the system. On compatibility, an SDK would allow people to build their own products that leverage the compute of the network. That's not something there is a high incentive for centralized render farms to offer. If you want to go deep I highly recommend recent podcast with CEO of OTOY and the two hosts of the Mograph podcast: https://www.youtube.com/watch?v=SRtOMF4JKd4 https://www.youtube.com/watch?v=SRtOMF4JKd4

f00zz 5y ago

From your link, I found this a bit odd: > As video games get more photo realistic (...) the demand for rendering is going to grow immensely. I don't see what a decentralized render farm has to do with video games, it's never going to render at real-time rates.

JayStavis 5y ago
Agreed that is one of the hairier challenges a concept like this faces. I do think that argument is worth unpacking though. Would a render node in your neighborhood have acceptable latency? IIRC they have demoed "real-time" NFT's based on an Unreal Engine node streaming frames at acceptable e2e latency. Personally I'd think a bit broader than frame-streaming for achieving "console in the room" levels of video game fidelity.
- duskwuff 5y ago
  The primary issue wouldn't be latency. It'd be access to assets (which can easily be tens to hundreds of gigabytes for a modern game) and upstream bandwidth (which is often as little as 5 Mbps for residential users). These aren't problems that a blockchain can solve. In fact, I'll go a step further and say: distributed graphics rendering, or other computation, isn't a problem which is made any easier at all by the application of a blockchain. It's yet another case of cryptocurrency enthusiasts taking a problem that people seem interested in and claiming, without justification, that The Blockchain Solves This.