Deadline 5.1 / 6.0 Wishlist

Deprecated Advisory Board Archive
1

Hi All,

Thought it about time I did another one of my mega wish-lists! Actually before I do that, I looked through my last wish-list from July 2010 and a good number of these requests have now been developed. Awesome! Anyway, onward to my ideal wish-list (ignoring what has already been announced with Shotgun & Draft in v5.1):

Priority (1=high, 3=low):

  1. “Maintenance Plugin” Job Type. A new type of Deadline job which executes only on a whitelist of slaves BUT for each slave that completes successfully, it blacklists itself from ever running on that slave again (unless re-queued). I would like to submit a job to X machines and execute a script, python, vbs, fire a executable, anything to each of these machines. This has been asked for a few times now and is now my most desired feature request.

  2. Renderer DBR Plugin system. Develop a system whereby Deadline can support Mental Ray / VRay DBR across the slaves.

  3. SMART Farming - “AI” - auto render wrangling. different levels of automation. Global switch to set level of wrangling automation? Auto sense fast jobs and re-queue remaining tasks into chunks and assign to free/fast spec machines. Intelligent distribution of certain jobs based on software license’s availability.

No Priority (just numbered for convenience):

  1. Enhanced Pop-up Handling Framework. Ability to interrogate ALL the controls of actual popup dialog as they appear. This would allow support to monitor “sub” jobs such as FumeFX simulations and update the Task Status Progress correctly. A job sitting at 0% progress after 10 hrs isn’t very helpful for feedback! Better “Sims” based job handling. Realflow, FumeFX, Naiad + add Houndini sim job support, including distributed bucket support?

  2. Dynamic Limit Groups for dynamic licensing flexlm systems. I’ve looked into this briefly and with any “flexlm” based licensing system, its possible to query the license server, find out how many licenses of a certain type are already checked out and then for the clever bit, dynamically increase/decrease the license limit group thereby allowing studios to maximise on their software license count. Add an additional variable as an override, allowing a certain number of licenses to ALWAYS be available, ie: not be hogged by the farm. You could have 2 types of limit groups in the future, “static” and “dynamic”.

  3. Concurrent jobs. Chad C’s idea. Use any wasted CPU cycles by allowing slaves to receive additional jobs to process on the ‘other’ cores. Lots of re-factoring of Deadline to do I guess, but amazing potential for efficiency gains? Need way to distinguish between “lightweight” and “heavy” jobs.

  4. Colourized Messages in the Log Reports. Now we have themes, ‘colourising’ the text in error/log/re-queue messages? ie: “Error” (red), “Warning” (yellow), “Information” (green), and “Debug” (grey). Coloured in the slave window and Pulse window as well? Tie it into colour themes, so user can control colour scheme of log reports windows! As a default, I would suggest leaving all INFO messages as they are, but colour up WARNING as yellow/amber and ERROR as red. This would make it really easy to see these important messages in the log/error/re-queue reports, as I think they tend to get missed sometimes…

  5. Search History in the search windows of the deadline monitor & repository. Global repository setting controlling number of search results to display. I see this as a drop-down list in each search history field in Deadline which simply remembers my previous X number of searches. I find myself carrying out the same searches all the time. This would make it quicker and also act as a reminder.

  6. Post image file validation - auto check that I don’t have any < x Kb files created. Therefore, corrupt, re-queue for me automatically. Ability to check individual channels/render passes and re-queue automatically would be mega-awesome. Also, ability to auto-calculate what the average frame size should be and therefore know that a certain frame doesn’t meet that standard and hence re-queue automatically.

  7. Output image thumbnails as an additional tab in the Jobs list section of the Deadline monitor. Thumbnail size controllable. Gives user basic image validation & optional image (passes) channel support? Right-click for more info on the actual thumbnails; resolution, channels, gamma, colour profile, time-code, etc.

  8. Developer IDE Intellisense Extension / Plugin for Visual Studio and/or Eclipse IDE

  9. Pulse Redundancy - Multiple Pulse instances for redundancy. Tricky I understand, but at the moment, the only way I can build redundancy into Pulse is by having it as a VM and if the VM or ESX host fails, then we have ESX host failover to allow it within 30 seconds to be re-hosted.

  10. Dependency Node Graph in monitor. Filter jobs by upstream/downstream dependency together with a filtering system (drop-down lists) like the monitor to allow visual review. Colour coded depending on state of each job in a certain dependency chain.

  11. Deadline Mobile - Control actual jobs in the queue? iPad edition, allowing control and graphing/charting?

  12. SQL database backend.

  13. Industry Certification - DCA - Deadline Certified Administrator Program, DCE - Certified Engineer/Developer for Deadline. Industry standard. More industry recognition? Perceived maturity of the product by potential clients? More likely to consider it than competitors?

  14. Video tutorials. Fact. Your average user doesn’t read the software manual. Reduce the number of ‘simple’ questions appearing on the support forum on basic Deadline setup. How about video tutorials of Deadline setup, config, app submission scripts, etc. Links in the Deadline wiki/online published manual to a hosted video site? Vimeo?

  15. Global farm reporting system. Ability to generate reports, almost like a SQL query. Fully customisable. ie: If I get asked to generate a report based on certain jobs all for a certain project. “How many rendering hours were used/wasted on project X” type of question. Ability to pull out images based on the graphing would be most useful for management. Helps me justify more render nodes, which equals more Deadline licenses!

Thanks for reading! Thoughts?
Mike

May 30, 2011 | Mike Owen

thanks mike - please keep in mind that anything i say doesnt mean it will make it into deadline in short order!

  1. is this to patch/modify install things? i ask because we are working on this in a different fashion.
  2. thats not on the roadmap for 5.1 at the moment, thats for sure.
  3. these things are starting to come together - we are big on autoconfigure and push-button stuff now-a-days :stuck_out_tongue_winking_eye: so this is a good suggestion that is similar to ideas we have bandied about.
  4. ryan can address
  5. my problem is not everyone is flex, and in fact, some are evaluating other solutions. if there was a generic api that multiple license companies reviewed, this would be higher priority for me [part of auto-config] but again - lots of fish to fry: not sure licensing is near term
  6. refactoring: not necessarily. the “easy” way for us and the $$ way [for you] is to allow multiple slave instances on a machine. would that be of interest? would you buy some more slaves to do this on your high powered machines? since the software is a small fraction of the cost of these computers anyway…?
  7. ryan!
  8. ryan?!
  9. thats been discussed. i want more of this kind of thing…i suggested using draft to diff between prevous and next frame renders and placing a bar beside the frames in deadline which denotes as follows: green bar that at 0-30% is red [denoting a flash frame, or possible bad frame] and at 60-100% is green [denoting frames have similar colour components etc] thus helping to analyze potential bad frames. i would say on the roadmap…not sure when!> :_)
  10. should come with draft
    11.Ryan?
  11. Thats part of our auto-config/self heal plan. at least in my head. heck, i want the repository to self-heal and replicate if it goes down :wink:
    i think this stuff will get introduced over the next n releases.
  12. we have a node based library. hmm. interesting idea. ryan?
  13. its on MY wishlist, and has been for a while.
  14. oh boy. i guess i need to know why you are asking for this? we discussed it and decided it wasnt necessary - and our shotgun integration demonstrates that we can write anythign to and read from a database so…what do you need that you cant have deadline article?
  15. i’ll need your help on that - eg. the community.
  16. already in progress. i have some, audio isnt great, working on improving the process for making these. btw, did you see the first deadline tutorial? we are going to be doing more in this space. video, tutorials, training…tips, tricks, stupid pet monkey - again, here - i reall y need your help. i would love to have you guys to do some tutorials. record em, script them, and i can have th em dubbed. or dub them yoruselves if you have a good voice… and proper recording kit! in all seriousness - i will happily pay some paltry sum to have some made by you guys.
  17. thats a part of our wishlist!

those are my thoughts…coutner arguments >? ;-PP

cb

June 1, 2011 | Thinkbox Software

  1. That’s been in the wishlist since 2.0. It’s not so much about MR or Vray, it’s about having a “task” that doesn’t just render and complete. The task just runs any application or service and when the job is suspended or a timeout is reached then the application closes. You could have a ‘webserver’ job or a ‘bittorrent’ job or a ‘bitcoin mining’ job or even a ‘Backburner’ job.
    This ties into our wish for a ‘workstation’ mode for slaves some.

June 1, 2011 | Chad Capeland
edit remove
Hey Mike,

No fancy first paragraphs. Might as well jump right in. :slight_smile:

  1. There are many uses for this feature which you have listed, and another one would be sanity testing scene files to check for discrepancies between render nodes before ramping up. The idea of using blacklists/whitelists seems to be the most reasonable way to fit this type of job into our current architecture, but maybe there is a better way…

Anyways, it’s something we continue to discuss internally, and while the roadmap for 5.1 is already finalized, we will make a note to discuss this when planning the 5.2 roadmap. That doesn’t guarantee anything, but I think it’s safe to assume that a lot of people (us included) like this feature!

  1. See Chris’ response. This just continues to remain a low priority feature, not because we don’t like it, but because there always seems to be more important features ahead of it. Plus, the Tile Rendering feature is already a good replacement in a lot of cases, and fits better with Deadline’s idea of job scheduling (as opposed to machine reserving).

  2. Some AI is definitely good, and as Chris mentioned, the Auto Configuration feature set is a big focus of ours and we have ideas to continue making it better (auto group assignment, and extending it to the Monitor, which you had suggested before). I know Qube has the ability to break up remaining tasks of a job into separate chunks and assign them to faster machines, but I’m not sure how big the impact of this is in the grand scheme of things.

For example, breaking up the last few chunks is really only useful if your farm is sitting idle. Besides, chunking is only ever recommended if a frame task less than a minute or two to render. Yes, there will be some edge cases where an artist makes a mistake and sets the wrong chunk size, but they could also just as easily submit to the wrong pool…

  1. Still on the wish list, but no ETA. If only these sim apps printed progress to stdout, or 3dsmax’s logging system, rather than use a UI element to do so…

  2. I think Chris summed this up well!

  3. I think multiple slave instances would be ideal for this. You could configure them to be in different pools/groups, and thus can handle different jobs (one gets a heavy job, one gets a light job).

  4. Another low priority feature that we would love to have.

  5. Good idea! I’ll add it to the wish list.

  6. Chris covered this.

  7. I don’t think there is much out there for embedded image viewers in .NET. I found this (mastropaolo.com/devildotnet/), but I’m not sure if it is useful, or Mono-compatible. We already allow the double-click option to open images in the application of your choosing, so why reinvent the wheel? Chris’ idea of image comparison is a different story though, and once Draft can do that, highlighting problematic frames should be simple to support.

Mind you, I guess if we converted images to JPEGs or PNGs for the thumbnails, .NET would have no problem displaying those.

  1. You know how bad I want this! It would probably be some sort of debug mode inside Deadline, rather than an IDE plugin, but that could change. This is low priority though, and isn’t on the current road map. :frowning:

  2. Yeah, that would be part of the self-heal stuff Chris was talking about. If a slave detects Pulse is down, it could start up Pulse on itself (or a backup machine) and then modify the repository settings so that all machines recognize the change. We would need to make sure we avoid false positives, but definitley something we would like to see.

  3. We would need a managed .NET node library, which I don’t think we have. Custom drawing a simple node based system shouldn’t be overly complex though. A treeview could accomplish this as well, but it would have to flag circular dependencies so that it doesn’t expand indefinitely. We already have been looking at the possibility of replacing our current job list control with a hybrid tree/list control…

  4. On the wish list, and I know Chris would love to see the feature set expanded, at least for the iPad version. :slight_smile:

  5. Database doesn’t always mean better… like Chris mentioned, we would need more information about why you would want a DB. We already have plans to store Deadline job stats in a DB…

  6. Covered by Chris.

  7. Covered by Chris.

  8. Once we get the stats into the DB, this should be easy.

Cheers,

  • Ryan

June 2, 2011 | Ryan Russell

  1. Tile rendering has a LOT of latency. We’re talking minutes instead of seconds, and for very large scenes it can be much worse. And that only covers the cases where you are emulating a bucket dispatch system. For something like reality server, it breaks down. And for anything not render related, it’s even worse.

June 2, 2011 | Chad Capeland

I am very interested in Deadline as a remote render farm host software as well eventually even deploying an image with everything that is needed to render the job. In connection to that i wonder how it could run on virtual machines (server and slaves) via VMWare and the like. Taking it to an isolated cloud pretty much (to make it secure to fall into NDA save land etc… ) like a bubble/sphere on a bigger cluster of machines. I have to admit i know more what i would want rather then having the technical IT background so please excuse in advance for rural outlines of ideas!

Scenario:

  • I book 50 virtual machines on a HPC(High performance computing) cluster for remote rendering/caching PRTs (lets assume i have access to HPC)
  • I secure the booked cluster with VSphere or other security layers (thats mostly up to the provider i’d think) to a degree that is government level security to fulfill NDAs and other agreements with the client i render for
  • I deploy a gold image of the config i need to tackle the task (OS, 3D App render slave, Deadline/Pulse, Plugins, etc.)
  • I run deadline on the cluster for the time i booked it (rental liceses/bought licenses)
  • Deadline “at home” can surveillance the progress happening on the remote cluster and intervene (cancel/restart job, etc. the full capabilities)

Or even better:

  • Same booked cluster as mentioned above
  • very basic gold image is deployed with deadline installed
  • Deadline sends first render job > checks if it has everything it needs (plugins, 3D App version, etc.) > if not first job calls “home” requests the plugins etc., essential a complete image with everything > deploys the image > renders/caches > last job looks around: “oh I’m last, we don’t need this particular image anymore” > initializes rolling back to basic image

I don’t know if it makes sense or is feasible but when Deadline is having a pick-nick on a foreign lawn it should bring what it needs to have a good time, if that makes any sense :slight_smile:

I like Mike’s ideas! Especially #6

Another thing, does the remote viewer (VNC, TeamView, etc.) work with LogMeIn over the web? I have not tested that but like team viewer it’s free and comes with an iPhone app. I use it here in the PRO2/Ignition version and it’s pretty convenient to check if everything is OK “at home”.

June 5, 2011 | Anselm v. Seherr - Thoss

Hi Ryan/Chris et al,

Prepare yourself…Here’s my counter arguments :slight_smile:

  1. I think with the new API functions that have been exposed in v5.0, this should be fully possible now? Its just a python scripting job, No? Another way Ryan? Interesting…tell me more!!
  2. Tile rendering is very useful and a good, core feature, but its just not the same when you’re studio is maxed out at the test lighting/rendering stage and very few final frames are going through the farm. I would much rather see a software toolkit, where I can rely on Deadline to assist my pipeline at different stages of a project. Whether that be file/data translation, DBR lighting/rendering, effects/simulation, to actual tile/final rendering, to finally, post processing, event driven dailies, etc. It would make for a much more rounded product if it could offer advatanages at all these stages (BTW - I think you already have most of these areas licked…except…). I feel, DBR is big news, whether its CPU, GPU or combo based. Unfortunately, I have no framework to capitlise on these tech advanatges, without having to step outside of Deadline to 3rd party tools, which is what I don’t want to do-( I’m sure bending Deadline’s architecture to handling DBR isn’t easy, but I do think its something that needs to be considered.
  3. Good point on chunking/de-chunking. Enhanced auto-config would be a good addition. Auto group, pool, monitor settings are obvious places to go.
  4. Annoying. I understand. I was thinking of more enhancement on the Deadline UI .NET dialog scanning front. ie: a 3rd party pop-up dialog appears, its all classed as static buttons, great…not much use, but what if you combine it with some regex action looking for matching UI element names and then grab the info required? OK, so it breaks potentially for each new version of a particular sim plugin when it comes out, but don’t we tend to confirm the compatibility of a version of a plugin as a version of Deadline ships? ie: Krakatoa 1.x now supports FumeFX version 2.x, kind of thing?
  5. Fair enough. Still worth keeping an eye on for the future. Say out of 40+ licensing systems I have on multiple floating license servers; I think at least 50% of them are Flexlm based. You can get Nagios Flexlm enviroment monitoring plugins to test license counts and report back via a web server…I was kind of thinking of a similiar setup, but hooked up to a ‘special’ kind of dynamic limit group.
  6. Multiple slave instances. Interesting idea. Solve’s CB’s issue with large core counts = low Deadline license count revenue. Licnese Deadline per number of instances? As long as each slave instance is effectively NOT connected to another slave instance for stability. ie: don’t want a house of cards issue, when just 1 of the instances goes down. I see this like Google Chrome, each tab in the browser is sand-boxed and if it gets into trouble, its lock-down and auto-recover time?
  7. Oh go on…a splash of colour in the log reports would brighten my day up!
  8. Cool. Search History would be useful…me thinks…
  9. Oh…I like CB’s idea. Nice. I’m thinking the initial basic checking of frames, data size, etc, could be achieved with just some scripting…shall I add that to my list of projects?! :slight_smile:
  10. I got the idea of basic thumbnails as an additional tab in the frames list window from this: lonerobot.net/?p=1003 Just thought it would be nice as a quick checker as you scan your frames. Beats having to double-guess which frame number you ‘really’ want to open up in hte external vieweer of your choice.
  11. Devleoper IDE extension. DEbug mode sounds useful. Hmmm…I may have a solution to this already…Eclipse IDE gives good code completion based on the existing py files open. Gives templates (snippets), code completion options and is cross-platform as well as open-source. How about whenever Deadline goes through a automated build, it has a post-build process which pumps out a XML file containing all the Deadline scripting API functions with integrated help desciptions? Then just plug this into Eclipse IDe via its pyDev plugin? Much easier to acomplish than a VS extension?
  12. Pulse reduncancy. I reckon quite a few studios rely on WOL / Power manaagment via Pulse. What happens when Pulse goes down? Failover to finally boot up on the repository sounds good after maybe trying to boot up on a few pre-defined slaves?
  13. “current job list control with a hybrid tree/list control” - this actually sounds brilliant. when you have lots of jobs in the queue, it can become a bit of a minefield. This would simplify things, but also make things more powerful with nodal visualisation, ie: dependency graph. Expandable/Collapseable tree control. Yeah, monitor doesn’t need to be full-screen anymore to see what the hell is going on!
  14. iPad app…Hmm, it is the best tablet out there at the moment, but probably a bit of a gimmick? I have to admit, the Qube! iPad app looked ‘well sexy like’ but it was only displaying stats pulled from its DB and squirted through a professional graphics template: roambi.com/ - Licensing for developers looked pricey last time I checked. Probably better things to spend your cash on.
  15. Stats in DB for Deadline is excellent. I think Ryan mentioned that he was seeing some nice performance via a DB. I’m simply asking if this is something that should still be considered for the main Deadline engine? No reason on my part, but I do look at the competition and like to compare notes :slight_smile:
  16. Cool no worries :slight_smile:
  17. Yep, I have a list I’ve made somewhere already of ‘starter’ video tutorials…
  18. Generating reports based on info from DB. Cool. But I’m alos thinking it would be relatively easy to pull this existing info via py scriping. Ah…bugger…I’ve done it again. I’ll add it to the project list :slight_smile:

Anselm - I believe Deadline is already supported in the VM enviroment (depends on exact OS, etc), but generally works fine. TB use VM’s all the time to test stuff? As Deadline is supported in the VM enviroment, HPC / cloud should work fine? Not sure about “LogMeIn”, but if you successfully establish a tunnel to your computer / network, then RDC / VNC all work fine.

Cheers,
Mike

June 6, 2011 | Mike Owen

  1. It needs to be a special job type. I’m sure it could be hacked together with the current API, but there could be race conditions and other issues, particularly when updating the “blacklist”.

  2. The biggest hurdle with DBR is that everyone does it different (vray, mentalray, etc), and there aren’t many options for a 3rd party manager like Deadline to control the DBR process. During some initial testing, we had something sorta working with vray, but it wasn’t a dynamic system. Vrayspawners couldn’t be added to the render on the fly, so you needed to wait until all your spawners connected before starting the render. This waiting game doesn’t work well in the context of Deadline, because it means machines are idling while the job is waiting for the last few slaves to join in.

I had send Vlado the list of things we needed to properly support Vray DBR in Deadline, and he made it sound like everything was reasonable and doable, but after that those talks sorta dropped off. We haven’t even looked at what needs to be done to control mentalray DBR yet. I’m thinking this sort of feature would require a big push on our part to get the developers of DBR systems on board and help us by adding new features to their systems.

  1. We rarely test Deadline against a specific plugin version, and instead focus on supporting specific versions of applications. This is because Deadline rarely interacts with plugins directly, and it’s basically an assumption that if Deadline supports 3dsmax 2012 (for example), it should support every version of a plugin that works with 2012. Krakatoa is a bit different, because it can interact directly with FumeFX.

It’s enough work keeping up to date with all the various applicatoins we support. Could you imagine us doing it for every plugin as well!? :slight_smile:

While I think advanced popup handling is a very cool feature, it’s scope is pretty limited for the amount of work that needs to be put in. The only mainstream use for it that I can recall is for pulling progress out of FumeFX, whereas the current popup handling functionality is probably used by at least 1/3 of our plugins. That’s why it tends to stay low on the priority list.

  1. Each slave would be completely independent. In fact, you would see each slave application running as a separate instance, and they could be launched and shut down independently of each other.

  2. Some viewers support this flipbook functionality, which can help take the guessing out of which frame to choose. Heck, even Windows explorer has a thumbnail view. :slight_smile:

  3. That’s an interesting idea. We’ll definitely consider it if/when we take a look at having proper IDE extensions. We would probably want to look at VS first, since that’s our main dev tool, but if all we need is an xml file for Eclipse, we should give it strong consideration as well.

  4. Currently, if pulse goes down, the only feature of power management that will still function is the Thermal Shutdown. I like the idea of having a backup list of pulse machines that could be an arbitrary length.

  5. So far, every hurdle we’ve hit that we though would require a DB to fix, we’ve fixed without using a DB. Maybe one day it will be inevitable, but when we get there, it will almost certainly be a complete backend swap out. In other words, there won’t be separate filesystem and DB-based versions. There would simply be too many versions to maintain.

Deadline works just fine inside of VMs, but I think Anselm was asking “what if Deadline rolled out the VM images prior to rendering on them?”. It’s something we’re looking at. :slight_smile:

Cheers,

  • Ryan

June 8, 2011 | Ryan Russell

  1. Why would you need special support? The “job” just starts another process. and when you end the job, it closes the process. If you wanted to watch something to see if it “completed”, you would just have the slave monitor the process you launched, and if it closed you dispatch a new job to the slave.

This would add “Deadline support” to any application. Maybe it’s not very intelligent or able to make decent reports, but it is universal. And if slave monitors CPU usage, you could even assign another job, so efficiency wouldn’t be impacted badly.

Regarding Anselm’s VMs, there’s a course at SIGGraph this year on LLVM rendering. It looks really interesting.

June 8, 2011 | Chad Capeland
edit remove
If that’s all you want to do, then you can already do that with Deadline. The Proof of Concept plugin I created for VRay DBR simply launched the vrayspawner app and let it run indefinitely. Then when the DBR render finished, you could just suspend or delete the job to allow the slave to move on. There are also API functions to have a Deadline plugin monitor a process, and you can override the callback that gets called when the process exits outside of Deadline control. This would allow you to mark the task as successful, rather then the default behavior of failing the task.

If we’re going to support something like Vray DBR out of the box, I would want to do it right. I want to see slaves joining the DBR process on the fly so that a DBR job can start as soon as one machine picks it up, rather than waiting for all of them to join. Otherwise, any slaves that try to join after the fact will just sit their idling. This way, a DBR job fits into the queuing system just like any other job.

Cheers,

  • Ryan

June 8, 2011 | Ryan Russell

Hi,
Just a few notes I’d like to add:

  1. Sounds brilliant. However, the slaves list will start to look very busy/complicated. So I suggest implementing the tree control at the same time. So you would have 1 tree control for the jobs list displaying dependencies! Cool! AND another expandable tree control to display the machines and their number of slave instances! Now, the clever UI bit, will be to clearly / cleanly display and have the ability to display which job is being processed on which particular machine AND on which particular slave instance of a machine. Colour coding and then something else would probably be quite good. How about little slave instance number icons OR “(x)” number cross-referenced to jobs/job frames next to each slave instance in the monitor?

  2. Yeah, I reckon the quick XML pump out route is a quick WIN for you guys to give better script/plugin DEV support. I’d write a tutorial up on this one if it all worked as well as I hope :slight_smile: In fact, you could even write “templates” for std lines in Deadline scripting such as declaring all the popular UI elements. I might even be able to knock up a proof of concept shortly…

  3. Pulse backup list - very cool / useful.

Cheers,
Mike

June 9, 2011 | Mike Owen

Deadline works just fine inside of VMs, but I think Anselm was asking “what if Deadline rolled out the VM images prior to rendering on them?”. “It’s something we’re looking at. :slight_smile:” <<< SWEET!

“Regarding Anselm’s VMs, there’s a course at SIGGraph this year on LLVM rendering. It looks really interesting.” <<< Hope there will be video recording…wouldn’t make it this year to Siggy :frowning:

June 9, 2011 | Anselm v. Seherr - Thoss

  1. We had thought about using a tree/list control to group the slaves together, but this custom control will require a lot of work, and I don’t like the idea of holding up the multi-slave feature for it. Our current plans is to allow you to arbitrarily name your slave(s) so that they’re easy to distinguish in the Monitor (each slave will appear as a separate row). Eventually, the tree/list control will allow us to group them, but custom naming and separate rows should work in the interim.

Cheers,

  • Ryan

June 9, 2011 | Ryan Russell

  1. Hmm…I’m concerned, but we’ll see…That’s going to make quite a few rows for us! I hope there’s going to be a good filtering system or I might get RSI using the middle mouse button to scroll down the list!

June 9, 2011 | Mike Owen

Another thing i’d kill to have is a RenderPassVoodoo built in. I don’t wanna go into too much detail on what it is (that falls into NDA land i belive) but it’s a render pass manager and file name convention tool that Bobo wrote for Avatar back in the day. It made out lifes sooo much easier!
I miss it on an every day base :smiley:
I know that there is RenderPassManager(lukashi.com/LPM.php) but imo RPM is too convoluted these days and therefor it stands in it’s own way. RPvoodoo was the best stab at a solution for render passes, cameras and naming conventions that i have seen to the day. small, handy, fast, awesome.

Ansi

June 14, 2011 | Anselm v. Seherr - Thoss

Agreed. RPM is insanely complicated. Maya render/display layers OR Blur’s Render Pass Manager are probably the best attempts I have seen in my travels…Of course I have never seen Voodoo tho…
What stops Bobo from re-writing Voodoo again for Thinkbox in the future? :slight_smile:

June 14, 2011 | Mike Owen

“What stops Bobo from re-writing Voodoo again for Thinkbox in the future? :slight_smile:”

i really hope nothing! :smiley:

June 16, 2011 | Anselm v. Seherr - Thoss