General GPU/testing discussion from benchmark thread

145791018

Comments

  • neumi1337neumi1337 Posts: 18

    Just for info and if anybody want to test it. The Iray Server on irayplugins.com had been updated to 2.53 which contains Iray RTX 2019.1.1, build 317500.2663 laugh

  • outrider42outrider42 Posts: 3,679

    It would have made sense for Daz to wait just a bit longer on the general release so that Iray 2019 could be a part of the public release rather than beta. After all, it had been a very long time since 4.10 had updated, why not wait just a bit longer? But they sure did work fast to get a general release out in time for Daz strand hair.

    So how long will we be sitting on beta yet again to use the latest Iray?
  • RayDAntRayDAnt Posts: 1,120
    edited June 2019

    It would have made sense for Daz to wait just a bit longer on the general release so that Iray 2019 could be a part of the public release rather than beta. After all, it had been a very long time since 4.10 had updated, why not wait just a bit longer? But they sure did work fast to get a general release out in time for Daz strand hair.

    So how long will we be sitting on beta yet again to use the latest Iray?

    Strand Hair has probably been THE top secret feature in development over at Daz for quite some time. And remember that RTX was an Nvidia top secret project up until last summer - so far as Daz knew there were no new Iray rendering innovations to work into their release sequence until then. And as I understand it, Daz has been sitting on the IP underpinnings to Strand hair for 2+ years. Hence the seemingly underfeatured 4.11 release. My guess is that we may see a subversion incremented 4.11 rerelease in the near future (it's happened at least once before iirc) or possibly a REALLY fast jump to 4.12.

    Post edited by RayDAnt on
  • Saxa -- SDSaxa -- SD Posts: 871
    edited June 2019

    It would have made sense for Daz to wait just a bit longer on the general release so that Iray 2019 could be a part of the public release rather than beta. After all, it had been a very long time since 4.10 had updated, why not wait just a bit longer? But they sure did work fast to get a general release out in time for Daz strand hair. <.<

    So how long will we be sitting on beta yet again to use the latest Iray?

    Have you had bad luck with Betas?

    I was updating with Beta since I started in February of this year, so i don't have the length or depth of experience as many of you do.  But I was delighted to find that Beta worked well for my many test usages.  Was it that some things did not work well?

    To me it makes sense that they rolled out Strand hair in General Public Release State for those that want to play with Public Iteration #1 of SBH/DF-SBH.

    And now they work on implementing RTX which may take ?  Plus they seem to be focusing on combining the speed increase of Ray tracing with animation improvements, which also makes sense.  As Daz-steve rumored about late summer for an animation release update, which may have delays due to unforseen this or that?  That's a longish time for not releasing General DS release with SBH support, especially with first DF-SBH hairs available for testing the market.  Daz-teams seems to have some pretty smart-cookies working over there, so am guessing they have a good game-plan.

    Post edited by Saxa -- SD on
  • outrider42outrider42 Posts: 3,679
    Yes, I used the beta. But a lot of people don't use beta software for different reasons. We had a user recently who builds for a company and they will not allow beta software, period. Also a number of plugins don't work in the beta without doing some work arounds. Plus there are people who don't even know how to get the beta! And last, I think its just poor form to ask your customers to use a beta for an extended period of time. 4.10 dates back to 2017, and only just NOW has it received an update. That's not cool. It doesn't matter in the slightest when Daz started working on dforce to me. Dforce strand hair only showed up a month ago in the beta, and boom, its now in the full release lickety split. Meanwhile other features like a functioning denoiser, Iray chromatic fixes, and updates to cloth dforce went on for over a year in beta without ever seeing a general release.

    Not to mention that Nvidia publically stated Iray 2019 would ship in May. I guess they missed that mark, but we all knew it was coming soon. Daz Steve said it was coming several months before that. He promised it was coming to Daz.

    If the full release gets updated to the new Iray quickly after it launches, then swell, but I'd rather not have a year+ wait for Daz to launch a full release again.
  • bluejauntebluejaunte Posts: 1,859

    Beta is just a word. There's Blender 2.8 right after alpha, that was the type of beta software you don't wanna use. Daz Studio is never in that state in the puplic builds. Daz Studio could literally be in beta state just because there is a new Iray version and nothing else changed, in theory anyway.

  • Robert FreiseRobert Freise Posts: 4,247

    Blener 2.8 is a beta

  • outrider42outrider42 Posts: 3,679

    Blender is also an open source software, that's not a proper comparison. Because of the way open source is built, its only natural for them to have long running betas, and nobody is forced to use them, plus they are free software (and don't try to counter that Daz is free, it is "freemium" software), and besides, you can build everything in Blender, but Daz does not allow customer to build HD models or strand hair with dforce hair...customers have to purchase them. That's the key here. We are Daz customers, we have to create an account just to use Daz Studio. But people using Blender are not Blender customers, you do not create an account to use Blender, nor do you buy assets and 3D products for Blender from a Blender store. The expectation is different. Blender does have a store, but it is not an asset store. They just have swag like clothing and such, though they do have tutorials in there, too.

    Blender also offers downloads of many of its previous versions, so if by chance something is wrong or you simply do not like it you can go back to the version you prefer. The user always has the choice of which version to install. Daz does not offer any choice. You have the newest beta and the newest general release, and they are the only choices you have. If you update to the latest of either, you cannot go back. Period.

    Beta is not just a word. Betas are software that is in an unfinished state. That is why most companies will not use beta software, like the Daz user's company I spoke of. A beta is a serious no go to them. Just because the previous Daz beta was stable does not mean that all of them will be in the future. I've used other beta software that contained experimental features, and these features were very unpolished or unstable. That is why it is a beta. Betas are usually conducted to get feedback and testing so that the developer can improve the final product.

  • RayDAntRayDAnt Posts: 1,120
    edited June 2019
    Meanwhile other features like a functioning denoiser, Iray chromatic fixes, and updates to cloth dforce went on for over a year in beta without ever seeing a general release.

    To be fair, only one of these things (dForce) is something developed by Daz (same thing with RTX support.) And dForce updates have all been incremental. So calling them out for essentially not taking active credit for updates to software/ features not developed by them...

     

    Yes, I used the beta. But a lot of people don't use beta software for different reasons.

    Conflicts surrounding people's subjective standards for what constitues alpha/beta/gamma (release)/etc software and reasonable use (especially in business settings) has always been murky. And is destined to remain that way so long as people remain subjective. There's lots of extremely popular software out there which tends to spend much of its useful release life classified as beta software despite performance/reliability comparable to most other developer's "finished" release standards. And there's other software whose "releases" wouldn't meet the standards for most people's alphas (hello Microsoft...) It's annoying, but c'est la vie. Smart companies understand this and adjust versioning acceptability standards on a deveoper case-by-case basis. Less smart companies don't.

    Post edited by RayDAnt on
  • bluejauntebluejaunte Posts: 1,859

    There isn't a norm as to what is beta or not. Some software is stable as hell and in beta all the time because the developer can't be bothered to admit it's probably stable enough to be called final. None commercial software especially, since they don't need to sell anything they can literally leave the beta state forever so they won't have to feel as responsible when something does go wrong. Gmail was in beta for years and everyone was using it.

    Beta is a word that the developer can use however they want. It could be used for a new game that is barely past alpha and still has a very long time to get to a state that is even remotely usable, or it could be something like a Daz Studio beta that is really software that has been "finished" a long time ago but just got one new feature that might not work 100% yet, which does not mean that the whole software is suddenly unstable or not working as expected at all.

    I put finished in quotes because software is never finished. It is always in an unfinished state. Even software that is no longer developed is not finished. Bugs remain, half developed features remain, or features remain missing entirely.

  • outrider42outrider42 Posts: 3,679

    The term "beta" is not subjective. It is well defined. While one can argue the quality of software produced, one cannot argue what the term "beta" actually means. Here it is from Wikipedia. Take special note of the last paragraph.

    "Beta[edit]

    "Beta test" redirects here. For the film, see Beta Test (film).

    Beta, named after the second letter of the Greek alphabet, is the software development phase following alpha. Software in the beta stage is also known as betaware.[3] Beta phase generally begins when the software is feature complete but likely to contain a number of known or unknown bugs.[4] Software in the beta phase will generally have many more bugs in it than completed software, speed or performance issues, and may still cause crashes or data loss. The focus of beta testing is reducing impacts to users, often incorporating usability testing. The process of delivering a beta version to the users is called beta release and this is typically the first time that the software is available outside of the organization that developed it. Software beta releases can either be public or private, depending on whether they are openly available or only available to a limited audience. Beta version software is often useful for demonstrations and previews within an organization and to prospective customers. Some developers refer to this stage as a previewpreview releaseprototypetechnical preview / technology preview (TP),[5] or early access.

    Beta testers are people who actively report issues of beta software. They are usually customers or representatives of prospective customers of the organization that develops the software. Beta testers tend to volunteer their services free of charge but often receive versions of the product they test, discounts on the release version, or other incentives."

    It is the definition of beta that scares away people from using the beta. In particular, the possibility of data loss is frightening enough on its own, and this where many companies will say hell no to any suggestion of using a beta software on company computers. Most Daz users do not want to be beta testers! So with the definition of beta in mind, many people avoid the beta. You guys need to remember that many people here are not tech savy. A lot of people do not know the world of software like you do, and they will not have the confidence for installing a beta. 

    So no matter how amazing and stable the Daz beta may be, it is still called a beta and thus falls into the definition above. Besides, if the beta was so great, then why was it not released as a general release for over a year? Answer that question, please. When you have a general release from 2017, and then ask your customers to just use the beta for ALL of 2018, what? The general release did not even support RTX at all until last week, over half a year later. And as for the argument that Daz cannot control when Nvidia gets an Iray update out, well, too bad, Daz chose to use Iray in the first place. That decision is on them.

    Also, while Microsoft is gonna be Microsoft, they also allow customers to roll back Windows if an upgrade is not to their liking, unlike Daz. I really believe that is a major issue with Daz Studio, there is no way back from upgrading without planning ahead before Daz releases an update. Remember, we had a thread warning users that a full release appeared to be coming soon, and to back up their installs...why would we need this kind of thread? If something breaks, too bad. And Microsoft does not ask the average user to use the Windows beta. But here at Daz, if you bought an RTX card, you had no choice, and asking users to use the beta was the standard response.

    And there's another issue, getting the Daz beta is not straight forward at all. You cannot simply search for the beta in DIM. Nope, not that easy. You have to specifically uncheck a certain box in DIM to see hidden items, and this is something most people don't know. We still get a lot of people up in the forum who have no idea. Either users have no idea a beta of Daz even exists, or they do not know how to access it. Think of all the new users of Daz Studio who don't know this stuff. This is something I always consider. You guys have done this stuff for years, for once just try to imagine you are new to all this and don't know what is what. You are not going to find the beta on your own.

    I've dealt with long betas. I backed the video game "Next Car Game", which would eventually become "Wreckfest". This game was in beta for ages, so long I kind of forgot about it when it finally did release. The game is still updating, though slowly. But this was not smooth sailing. Being in beta for so long angered customers, who itched for a finished product. The game lacked many basic features of a proper game for a long time. I can just imagine if those angry players were Daz Studio users. You guys don't want that, I am quite tame compared to many of them, LOL.

    But all of this strays from what the actual difinition of beta is. It does not make sense to ask people to run beta software for as long as 4.10 was around. If Daz insists on running long term betas while the general release goes out of date, it will come back on them some day. This is a customer driven business, and it only makes sense to keep them happy. Unlike Pascal, we had RTX cards working on day one of their release, which was fantastic...but only in the beta. They could have punched out a general release just to make RTX users happier. They could put out a general release after the big update to dforce which made it so much more stable. But they didn't. There were many points in time where a general release would have made sense.

    But it was only after strand hair came along that Daz suddenly pushed out a general release, and yet Iray 2019 is right on the horizon.

  • outrider42outrider42 Posts: 3,679
    edited June 2019

    Anyway, we have Nvidia Super to look forward to soon. The announcement for exactly what Super is will be on July 2. Super was teased a while back, but nothing has been revealed since.

    Leaks and rumors suggest that Super is a small refresh of current RTX cards. The 2060, 2070 and 2080 will be getting "Super" editions. Of particular note is the 2060 Super, which is supposedly going to get a bump to 8GB of VRAM. But all of this is just speculation at this point. We shall see what Super really is on July 2, so stay tuned. Certainly hold off on buying a new GPU for the immediate future if you can until we can see just what Super is. Does it involve price drops? Possibly.

    A link to the current rumors.

    https://videocardz.com/81076/nvidia-to-announce-geforce-rtx-2080-super-2070-super-and-2060-super-on-july-2nd

    Post edited by outrider42 on
  • bluejauntebluejaunte Posts: 1,859

    Yeah 4.11 was in beta too long, I did not personally understand this move. Why not just release 4.11 without dForce hair and stick all of that into a separate 4.12? This case kind of shows exactly what I was saying though. Daz chose to keep this in beta for as long as they did. It was their choice entirely. At some point, the very same software that was called beta could have been released as final but they chose not to.

    There may be an accepted definition of what beta means, but when have such defintions ever done anything? Have you never seen a software that was "final" and crashed left and right? A game that was released way too early? If everyone would follow this clear defintion of beta, then such things wouldn't happen. But they do, all the time. Likewise, software is kept in beta state despite it working perfectly fine. Some companies use the beta tag to say "we're still developing features for this", or it might even be a marketing ploy to give a software a new and edgy spin. The term means very little unless everyone adheres to the same standards, but they don't.

  • RayDAntRayDAnt Posts: 1,120
    edited June 2019

    The term "beta" is not subjective. It is well defined.

    Yes, but there are multiple definitions of it - eg. here's the one favored by Daz (also from Wikipedia):

    Perpetual beta (or the 'banana principle') is the keeping of software or a system at the beta development stage for an extended or indefinite period of time. It is often used by developers when they continue to release new features that might not be fully tested. Perpetual beta software is not recommended for mission critical machines. However, many operational systems find this to be a much more rapid and agile approach to development, staging, and deployment.

     The nature of 3D rendering software especially is such that constant iterations on software/user interface development are needed - which is anethema to the concept of "finished" versions (as bluejaunte rightly puts it.) Hence all the beta'ing. And if Daz were to suddenly switch over to a more "finished" focuused release model, users would simply (rightly) be complaining about having to constantly update their supposedly finished software.

     

    Unlike Pascal, we had RTX cards working on day one of their release, which was fantastic...but only in the beta.

    Again, just to clarify, RTX support was only lacking in terms of Iray (developed by Nvidia) functionality. Daz Studio 4.10 (the only thing Daz makes/has coding control over) was itself fully RTX compatible from day one.

     

    Anyway, we have Nvidia Super to look forward to soon. The announcement for exactly what Super is will be on July 2. Super was teased a while back, but nothing has been revealed since.

    Back on May 5th, industry insiders let it slip (ohne Deutsch here) that Nvidia was about to discontinue their practice of pre-binning better performing GPU dies as "A" variants for all cut-down Turing chips due to major improvements in the manufacturing process making the practice no longer necessary (eg. virtually all RTX 2060 TU106 dies are now coming out as "A" level or better performers.) The existence of these "SUPER" cards is just Nvidia craftily capitalizing on this same dramatic uplift in manufacturing yields on the consumer level front-end in such a way as to muck with AMD's current releases - because tech giants love trolling each other.

    tl;dr: Unless you are in the market for a 2080ti or Titan RTX (since these two cards already feature fully spec'ed dies) DON'T buy any RTX cards until after SUPER fully releases. Because you will be getting significantly better raw performance for the same or less money on virtually all other (lower) variants.

     

    Post edited by RayDAnt on
  • outrider42outrider42 Posts: 3,679

    Even Daz themselves state outright that the Daz Studio beta SHOULD NOT be used in mission critical machines. You linked it yourself. End of debate. You can guys can argue all you want, but there are people who are effected by this. And once again, you guys assume that regular users are tech oriented enough to understand all of this. That regular users even KNOW there is a beta, as there is no notification of a beta outside the walls of this forum, and even then you will only find talk of the beta in specific places. There are a lot of people who barely know much more than how to power on their computer here. This is a forum of artists and people use Daz because it is generally more user friendly than other similar software, and you can buy a preset for just about anything.  We also have a lot of Apple users here! (This is a joke...maybe.)

    The irony about this is that I have seen far more people complain about the 4.11 full release than anybody ever complained about the betas...what is up with that? I would make a suggestion of my thoughts as to why that is, but I would get into trouble. But I can say this, that so many people are having issues points to how many people still used 4.10 over the beta, since the full release has many of the features that the beta has had for some time. So you can keep arguing about the quality of the beta, but we have solid proof that a fair number of people never used it.

  • outrider42outrider42 Posts: 3,679
    edited July 2019

    Moving on, the "Super" variants of RTX have been officially revealed, and various outlets are already benchmarking them for games and other software ahead of the release. Sorry, Iray is not on these benchmark lists.

    But here are the basics. The 2060, 2070, and 2080 are all getting Super varients. The 2070 Super and 2080 Super will completely replace the original 2070 and 2080 which will no longer be sold after stock runs out. The original 2060 will remain available as it bridges the gap for RTX. The details on the 2080 Super have yet to be revealed, and this card will be launched in a few weeks time after the 2060 and 2070 Supers.

    This handy table comes from anandtech. I will refer to the Super cards as "S" for short hand. It does not list RT and Tensor cores, which annoys me to no end, but the Super cards get a small bump there, too.

    NVIDIA GeForce Specification Comparison
      RTX 2080 RTX 2070 Super RTX 2070 RTX 2060 Super RTX 2060
    CUDA Cores 2944 2560 2304 2176 1920
    ROPs 64 64 64 64 48
    Core Clock 1515MHz 1605MHz 1410MHz 1470MHz 1365MHz
    Boost Clock 1710MHz 1770MHz 1620MHz 1650MHz 1680MHz
    Memory Clock 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6
    Memory Bus Width 256-bit 256-bit 256-bit 256-bit 192-bit
    VRAM 8GB 8GB 8GB 8GB 6GB
    Single Precision Perf. 10.1 TFLOPS 9.1 TFLOPS 7.5 TFLOPS 7.2 TFLOPS 6.5 TFLOPS
    TDP 215W 215W 175W 175W 160W
    GPU TU104 TU104 TU106 TU106 TU106
    Transistor Count 13.6B 13.6B 10.8B 10.8B 10.8B
    Architecture Turing Turing Turing Turing Turing
    Manufacturing Process TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN"
    Launch Date 09/20/2018 07/09/2019 10/17/2018 07/09/2019 1/15/2019
    Launch Price $699 $499 $499 $399 $349

     

    The 2060 S is quite interesting, as now it has 8GB like its beefier siblings do, while selling for $400. Gaming benchmarks show the 2060 S beating the 1080 easily in most games. And with 8GB it compares quite well to the 1080, and it gets even better if you consider the ray tracing and tensor cores. This will make buying the 2060 S much more compelling over the 1070ti or 1080 as we can no longer dock it for just having 6GB. The 2060 Super is now my recommended starter card for RTX goodness, being $50 more than the original 2060.

    The 2070 and 2080 Supers are simply modest bumps in core counts over the originals. The 2070 S is close to the original 2080 in performance. The 2080 S will be a step above that, but do not expect it to come close to the 2080ti.

    It is very likely we will see some deals on the older 2070 and 2080 as they need to be moved from inventory. Plus they are outclassed, so nobody can justify their original price now.

    Also, I forgot to mention that the Founder's Edition pricing has been dropped on these, so there is that.

    Post edited by outrider42 on
  • RayDAntRayDAnt Posts: 1,120
    edited July 2019

    Even Daz themselves state outright that the Daz Studio beta SHOULD NOT be used in mission critical machines

    Correction: that "SHOULD NOT..." bit is Wikipedia conjecture - not official Daz opinion on the matter. Lest I be accused of putting words in Daz's mouth.

    You can guys can argue all you want, but there are people who are effected by this. And once again, you guys assume that regular users are tech oriented enough to understand all of this. That regular users even KNOW there is a beta, as there is no notification of a beta outside the walls of this forum, and even then you will only find talk of the beta in specific places. There are a lot of people who barely know much more than how to power on their computer here. This is a forum of artists and people use Daz because it is generally more user friendly than other similar software, and you can buy a preset for just about anything.  We also have a lot of Apple users here! (This is a joke...maybe.)

    Imo Daz Studio sits in a pretty unique, tricky place so far as 3D graphics rendering solutions are concerned. It caters itself to amateur/professional artists with rudimentary understanding of the tech involved (imo a GREAT thing, since there are so few solutions out there that do that) but at the same time focuses on incorporating cutting edge (which is to say, under-tested/oft' times undocumented) technological advancements as active features (dForce comes to mind) and relies on thoroughly not beginner-friendly things like Iray for core pieces of functionality. The end result is an extremely powerful rendering platform that - while on the surface is much easier/useful to use for making pure art than its competitors - becomes significantly more complex/mysteries than those same competitors once you start looking under the hood.

    To really get your money's worth from it, you have to either be a tech luddite who just points and clicks to nirvana or a tech expert who understands how to reverse engineer things to do your bidding. And as for all the people out there who sit somewhere between those two extremes... here's to hoping that Daz's next project after Stranded Hair/RTX stuff gets ironed out is some thorough documentation. yes

    Post edited by RayDAnt on
  • RayDAntRayDAnt Posts: 1,120
    edited July 2019

    Moving on, the "Super" variants of RTX have been officially revealed, and various outlets are already benchmarking them for games and other software ahead of the release. Sorry, Iray is not on these benchmark lists.

    But here are the basics. The 2060, 2070, and 2080 are all getting Super varients. The 2070 Super and 2080 Super will completely replace the original 2070 and 2080 which will no longer be sold after stock runs out. The original 2060 will remain available as it bridges the gap for RTX. The details on the 2080 Super have yet to be revealed, and this card will be launched in a few weeks time after the 2060 and 2070 Supers.

    This handy table comes from anandtech. I will refer to the Super cards as "S" for short hand. It does not list RT and Tensor cores, which annoys me to no end, but the Super cards get a small bump there, too.

    NVIDIA GeForce Specification Comparison
      RTX 2080 RTX 2070 Super RTX 2070 RTX 2060 Super RTX 2060
    CUDA Cores 2944 2560 2304 2176 1920
    ROPs 64 64 64 64 48
    Core Clock 1515MHz 1605MHz 1410MHz 1470MHz 1365MHz
    Boost Clock 1710MHz 1770MHz 1620MHz 1650MHz 1680MHz
    Memory Clock 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6 14Gbps GDDR6
    Memory Bus Width 256-bit 256-bit 256-bit 256-bit 192-bit
    VRAM 8GB 8GB 8GB 8GB 6GB
    Single Precision Perf. 10.1 TFLOPS 9.1 TFLOPS 7.5 TFLOPS 7.2 TFLOPS 6.5 TFLOPS
    TDP 215W 215W 175W 175W 160W
    GPU TU104 TU104 TU106 TU106 TU106
    Transistor Count 13.6B 13.6B 10.8B 10.8B 10.8B
    Architecture Turing Turing Turing Turing Turing
    Manufacturing Process TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN" TSMC 12nm "FFN"
    Launch Date 09/20/2018 07/09/2019 10/17/2018 07/09/2019 1/15/2019
    Launch Price $699 $499 $499 $399 $349

     

    The 2060 S is quite interesting, as now it has 8GB like its beefier siblings do, while selling for $400. Gaming benchmarks show the 2060 S beating the 1080 easily in most games. And with 8GB it compares quite well to the 1080, and it gets even better if you consider the ray tracing and tensor cores. This will make buying the 2060 S much more compelling over the 1070ti or 1080 as we can no longer dock it for just having 6GB. The 2060 Super is now my recommended starter card for RTX goodness, being $50 more than the original 2060.

    The 2070 and 2080 Supers are simply modest bumps in core counts over the originals. The 2070 S is close to the original 2080 in performance. The 2080 S will be a step above that, but do not expect it to come close to the 2080ti.

    It is very likely we will see some deals on the older 2070 and 2080 as they need to be moved from inventory. Plus they are outclassed, so nobody can justify their original price now.

    Also, I forgot to mention that the Founder's Edition pricing has been dropped on these, so there is that.

    For what it's worth, you can calcate the total RT core, Tensor core, and SM count for any given Turing architecture based GPU by dividing its total Cuda core count by 64, 64 and 8 respectively. So to fill out the missing parts of that Anandtech chart:

    NVIDIA GeForce Specification Comparison
      RTX 2080 RTX 2070 Super RTX 2070 RTX 2060 Super RTX 2060
    SM Count 46 40 36 34 30
    RT Cores 46 40 36 34 30
    Tensor Cores 368 320 288 272 240

     

    Post edited by RayDAnt on
  • outrider42outrider42 Posts: 3,679

    I think a lot of forget about the people who are not so well versed on tech. Its important to remember that. Plus Daz gets new users all the time. Everyone comes from wildly different backgrounds, and a fair number of users have barely used a computer before. Some users are older people who have reached a point to where they have some income and time to spare on things like this. There are all kinds. So no matter how fantastic a beta software may be, a lot of people do not know it exists, don't know how to get it, or are afraid of the term "beta".

    I have to urge people to use the beta myself, and I have told people that the beta is safe. Even then some still don't download it. The word beta has a stigma to it.

    Its extremely unrealistic to expect a majority of users ever used the beta.

    And I bet the people at Daz would stand by that statement that the beta is not recommended for mission critical hardware, simply because of the unknown factor of what a beta may do.

  • bluejauntebluejaunte Posts: 1,859

    Unlikely such people will have RTX anytime soon, and so no need for the beta. Even if you have RTX, there is no need to use the beta just because it features the latest performance enhancements. It's still plenty fast without it. If people are in the know about such things and explicilty want the boost, they will find out about the beta and figure out how to use it. If not, they will simply get the enhancements a little later. I don't see what the big deal is and feel we should move on from this "beta" debate. Beside the semantics aspect I don't see that any of this will change how Daz handles their development practices.

  • ebergerlyebergerly Posts: 3,255
    edited July 2019

    IMO, the term "beta" is somewhat meaningless and irrelevant. Those of us who have been around for a while have seen even release-version software cause huge problems. Last year a release version software took over my computer and took all my RAM until it crashed. I thought it was a bad virus. And for those of us who have worked in software development, it's not suprising. As I always say, this stuff is EXTREMELY complicated. It involves very complex software and very complex hardware. So how you name the software version is irrelevant. What's important is its impact. And nobody knows it's true impact until the universe of users actually tries it.

    And BTW, that's why developers HAVE alpha and beta versions...cuz it's so complex and they can't be certain the software works with every hardware configuration. So software is pretty much assumed to have issues at the outset, and they're hoping users will stumble on the issues and report them.  

    One of the things I dislike about the internet is it tends to encourage folks to vastly oversimplify stuff and think they understand it, when in fact they don't really know what they don't know. This was made crystal clear in what I consider the RTX fiasco last year, where people thought "wow, now we have awesome tensor cores !!". Same as with the GPU VRAM pooling. The facts are that this stuff is insanely complicated, and requires many levels of hardware and software to work together, which takes a TON of work and a LOT of time to implement. 

    Personally, I think it's extremely difficult for anyone to be a true expert in computers. Because you not only need to understand software (ie, be a developer) but also understand hardware design and application. Few people (none that I know, definitely including myself) really have that depth of experience. Especially when you're talking about proprietary software that nobody has access to, but somehow everyone is an expert.   

    And IMO, at the end of the day, do we REALLY need most of these new features, or are they just fun and exciting? My first question when a beta or new version comes out is "Why? Do I really need this, and am I willing to risk having bad things happen?". Personally, I tend to let the rest of the universe deal with betas and release versions to iron out the bugs, unless I really need the features. Which I rarely do. 

    And at the end of the day also, it's 9 months after RTX hardware was released, and where are we? Just like VRAM pooling. 

    IMO, forget the betas, wait another 6 months, and see what shakes out. If you really need it, that is. 

    Post edited by ebergerly on
  • outrider42outrider42 Posts: 3,679
    You can say that *now* about RTX, but you couldn't say it when RTX released, and until now. How many threads did we have where people asked why their shiny new RTX cards would not render in 4.10? Plus 4.10 had a nasty chromatic ss glitch, which directly effected bluejaunte models. This existed from the very beginning of 4.10 back in 2017. Only now does the full release address this. So Iray improvements extend beyond RTX, and go back much longer.

    The beta also offered big improvements to dforce long before the full version, long before RTX was a thing. So there have been a large group of people who have dealt with dforce explosions for a solid year when they didn't have to. I wonder if this impacted dforce sales? Obviously dforce is popular, but could it have been even more popular if the full release addressed it much sooner???

    The result, some people may have avoided dforce altogether. I had tons of issues with dforce myself and I hadn't noticed that beta updates had improved it.

    So there is a lot more to the beta than what Iray may bring. There are lots of other improvements in the beta that people using the full version never saw until now. That is the issue with taking so long between betas.

    Some of you have been doing this for years. Some of you work in software...and that is exactly the problem with your arguments. You have forgotten what it was like to not be in the know. For you, the beta is nothing, because you have done this for so long. Its nearly impossible for you to understand where people with vastly less experience are. Talking with you here demonstrates this quite clearly.

    Most people see "beta" and turn away. Simple as that. Your experience is meaningless because it is not THEIR experience. And a lot of Daz users are blissfully unaware that a beta even exists. Most people using Daz Studio are not browsing its forums on a regular basis.

    Daz does not advertise its beta. And let me ask you this, if the beta is no big deal and everybody uses it, then why did Daz wait to release dforce strand hair products until the full release had strand hair???

    After all, if the beta is no big deal and we all use it, why not just release strand hair products and ask people to use them with the beta??? Strand hair products only launched the very day that the full release got updated for it.

    That strand hair products only released when the full version updated should tell you all you need to know about how betas are treated and perceived by Daz3D.
  • ebergerlyebergerly Posts: 3,255
    edited July 2019
    You can say that *now* about RTX, but you couldn't say it when RTX released, and until now.

    Well, yeah, and in September of last year someone did:

    "Yeah, if it was me I'd wait until 1Q of next year. " 

    And the response was boo'ing cuz I was being too negative. Turns out I wasn't negative enough. 

    Post edited by ebergerly on
  • bluejauntebluejaunte Posts: 1,859

    4.11 beta has been unusually long. I developed my shader thinking this would be final soon, and when it became clear that this was not the case I provided 4.10 materials which were included at release of every character that was affected. The bug was also limited enough (only with transparent backgrounds) that QA agreed to let me include both materials. I later switched back to the "normal" SSS way that also worked out of the box in 4.10 because I realized I could still get the desired look some other way.

    So this was really more on me than the fact that this was in beta so long. It was also one of those unfortunate edge cases where some elusive bug could influence product development. This doesn't usually happen. Still, like I said I do think they should have released 4.11 way sooner with just RTX support and that annoying bug fixed and then throw dForce hair into 4.12. This was a case where I think it would have been fairly crucial to have a release version that worked with the latest hardware as fast as possible. Instead they dragged it out forever.

    Now we have a different situation. The latest Iray will merely increase the performance a bit, we hope. There's nothing very crucial about it. And if we're back to a more reasonable release cycle, it might be a mere month or two (or three) after the beta when 4.12 will be final again.

    I guess one thing I don't understand when you say these betas are too long, but then also that beta is a clearly defined word and hence is not meant for production in the estimation of the developer, shouldn't you then also agree with that estimation and patiently wait for the release?

  • bluejauntebluejaunte Posts: 1,859
    ebergerly said:
    You can say that *now* about RTX, but you couldn't say it when RTX released, and until now.

    Well, yeah, and in September of last year someone did:

    "Yeah, if it was me I'd wait until 1Q of next year. " 

    And the response was boo'ing cuz I was being too negative. Turns out I wasn't negative enough. 

    As far as I remember, you were being negative about RTX in general, not the time frame. And you later came around and became excited about RTX yourself. Now it seems you're back to being negative and talking about an "RTX fiasco". Not sure what prompted this latest episode.

  • ebergerlyebergerly Posts: 3,255
    edited July 2019

    Umm, no.

    In fact I was the one who was trying to get folks to understand that it was more than just "tensor cores", but instead a complete re-design of existing architecture, which impacted a bunch of different areas (Physx, Flex, MDL, rasterizatiion, ray tracing, CUDA, NVLink, Optix, etc.). But understanding that there's tons software and API's behind all of that, and how mind boggling it was for me to even comprehend how they could pull that off (unless they'd been working on it for 10 years already). So that made me certain it would take a very long time to fully implement, especially considering the end-user software apps had to be re-written to take advantage of all of that. My big unknown was how much had already been accomplished (which nobody knew or knows), and the more I saw the more it looked like it was gonna take a lot longer, or at least be a step-by-step implementation over a long period. 

    Though at one point I did see some Iray results that showed more than the typical 30-40% improvement between generations, which made think it might be closer than I had guessed. Now I'm not so sure, since it now seems like a "sometime this year" kinda thing, though I haven't been keeping up.   

    Post edited by ebergerly on
  • bluejauntebluejaunte Posts: 1,859

    You're talking specifically about Iray now? Because other renderers and some games already have support for RT cores. Anyway... if you look on page 6, RayDAnt has already posted about indications for full RTX support in the changelog of the private build. 

    Apparently you're also back to ignoring the fact that a 2080 TI renders twice as fast as a 1080 TI, and so at roughly twice the price is the exact same bang for the buck even before any of that fancy stuff gets implemented. Why people need to understand how complex everything is and remain "cautious" is beyond me. People who have the money and are willing to spend it will do so. Everyone else will wait. Nobody needs saving from making a horrible mistake by buying one of these RTX "fiasco" cards. Not to mention, if it turned out to be such a huge mistake they could just sell the card to someone else. Not an incredibly risky endeavour if you ask me.

  • outrider42outrider42 Posts: 3,679

    A number of PAs were shipping out models that used chromatic ss, so it was a wide spread thing, not to pick on you or anything. I certainly do appreciate the efforts as your characters are among the best around. (Though I do wish there was a male bluejaunte model, or maybe a person of a different ethnicity).The main issue I take is that 4.10 stood still for so long. I think there really should have been a full release at least half way between between 2017 and now. There really should have been a release that contained the chromatic fix, and this could have been done when dforce got its first big update. Dforce was and obviously still is a big deal, so I find the lack of a full release to improve it for the general users very odd.

    The RTX cards all perform extremely well with Iray even without the ray tracing cores. The performance gained has been well documented right here, and the performance gains for Iray are much larger than the gains most other software, like games, saw. But RTX is poised to make those gains much greater when Iray 2019 comes out. That is going to be an exciting day.

    Gamers were not very impressed by the initial RTX launch, but that mostly due to the sticker shock pricing. However, content creators have been in love with RTX from day one, and this love grew several times over in true Grinch-like fashion whenever their software of choice adds full RT core support. People are totally over the moon with RTX in other places where RT core support has happened. Hopefully Daz Studio will see Iray 2019 soon, and I feel we are just a couple months away from that.

    The latest launch of RTX Super largely fixes many of the issues of the original RTX launch. First of all there is no Founder's Edition price. The original FE cards were all around $100 more than the MSRP of the base cards, so a $500 2070 was actually $600 for the FE. This stung, because the previous generation 1070 was $380 at launch. The price increase for a 70 series card was too much, and this was true across the board for all the tiers. The performance was indeed there, but the prices were just too high. The 2070 offered gaming performance similar to the outgoing 1080. While it is kind of normal to see the performance of the last generation tier come down to the next, the issue was the price of these two cards was about the same...so where was the actual improvement? Gamers did not want to pay more just for unproven RT cores.

    Fast forward to July 2, and the 2070 Super is set to release for $500, and the FE will be that price as well. Not only that, but the 2070 S will perform very close to the original 2080, so now that $500 finally beats the last gen 1080.

    The 2060 Super does a similar thing, as it approaches the original 2070 in performance but for $400, and it has 8GB of VRAM.

    A lot of people would agree that the Super lineup is the way that Nvidia SHOULD have launched RTX back in October, and you will find a lot of media outlets repeating that statement. Either way, the Super lineup is much stronger than it was before.

    I also found a very juicy bit of information that few might catch. Because the 2070 Super uses the same chip and board as the original 2080, the 2070 Super has Nvlink connectors! And GamersNexus got Nvidia to confirm that Nvlink is indeed working on the 2070 Super. So now we have a new class of cards capable of Nvlink use at a much cheaper price than before. This means that you can get two 2070 Supers with Nvlink and have about 16GB of pooled VRAM for the price of $1000 plus the price of Nvlink. I know some people will find this very interesting news indeed.

    That means you can beat a 2080ti in both VRAM and performance for about the same price. Of course, this all assumes you can get Nvlink to pool VRAM in Daz Studio, which at this point I have not seen anybody succeed in doing. Though we don't exactly have many people who have Nvlink here, and I am not convinced that some know what to look for. Again, it is important to note that the standard GPU monitoring software will not report this pooled VRAM correctly, so the only way to know is to actually test it with a scene too large for a single GPU.

    I wonder if anybody is willing to try it? 

  • ebergerlyebergerly Posts: 3,255
    edited July 2019

    So let's see...

    Despite all the hype, the 2080ti gives, effectively, for the last 9 months, zero net gain over the 1080ti (twice as fast, twice the price). And now, they're suddenly introducting yet another line of "Super" cards which might actually be a better deal. And at this point all we have is a promise of some still unknown degree of future improvement in the RTX line sometime this year. 

    Yeah, I think I'll wait another 6 months to see what shakes out. laugh

    Post edited by ebergerly on
  • bluejauntebluejaunte Posts: 1,859
    ebergerly said:

    So let's see...

    Despite all the hype, the 2080ti gives, effectively, for the last 9 months, zero net gain over the 1080ti (twice as fast, twice the price). And now, they're suddenly introducting yet another line of "Super" cards which might actually be a better deal. And at this point all we have is a promise of some still unknown degree of future improvement in the RTX line sometime this year. 

    Yeah, I think I'll wait another 6 months to see what shakes out. laugh

    Zero net gain? That sounds a little strange, seeing as it's twice as fast (at least for us Irayers). You can now have one card instead of two for the same performance. You could also have 2x 2080 TI instead of 4x 1080 TI. the latter usually would be impossible on most mainboards and tough to cool.

    With a 2080 TI I also got better performance in games and viewports of various software, as these usually only ever use the main graphics card. It's plenty net gain for me. Did I mention I all got it for about 0 bucks because I sold 2x 1080 TI to cover the cost? And all of this is before full RTX support for Iray?

    There was no 2080 TI Super announced, by the way.

This discussion has been closed.