L'Ancien Regime

https://medium.com/@EightyLevel/how-voxels-became-the-next-big-thing-4eb9665cd13a How Voxels Became ‘The Next Big Thing’ We’ve talked with the amazing guys behind Atomontage, trying to figure out if the voxels can actually return and take down the polygons. WHY NOW? Large scenes made out of millimeter-scale voxels presents a difficult data management problem Atomontage has focused fifteen years of R&D on this challenge, resulting in patent-pending breakthroughs that finally allow volumetric simulations to scale in deeply interactive 3D applications Other features recently proven viable with our voxel tech include complex rigid-body physics, soft-body deformations, standard PBR shaders, and efficient playback of multiple volumetric video streams Around the 34 min mark a questioner brings up 3D Coat...while asking them about down and up resing models using algorithms. In their talk they mention another voxel program; Magicavoxel https://ephtracy.github.io/

http://www.cgchannel.com/2019/02/quadspinner-ships-gaea/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+cgchannel%2FnHpU+(CG+Channel+-+Entertainment+Production+Art) Gaea is available for Windows 7+ only. The free Community Edition, which is licensed for commercial use, provides access to most of the key tools, but caps export resolution at 1K. The Indie edition, which caps export resolution at 4K, costs $99; the Professional and Enterprise editions, which provide unlimited resolution plus a range of advanced features, cost $199 and $299 respectively. https://medium.com/quadspinner/mini-tutorial-drawing-a-mountain-range-c14ec7eec722 Drawing an entire mountain range.

Cases are nice but there's the UPS there's a second rig, there's switches and routers and hard drive storage in one of those hot swap arrays to play with. It turns into a big mess on the table and it becomes a major project to keep it all dust free. A used tower is $250 to $350. And I'm seriously looking at virtualization. VMWare Server 15 is only about $115 right now. Run that on top of linux. The pass through to hardware like the GPU is so good now that the layer of virtualization only takes up about one CPU cycle, which is really negligible. I'm thinking of two Samsung 970 M2's at 500 GB in RAID 1 and then have everything else offloaded to virtual machines for VMWare stored on a hard drive outside my main server. There are other such options to look at.

See, 2933 MHz is the top speed supported by Ryzen+ and that G. Skill Flare X399 is ECC. That's why it's pricey. A waste of money? If you're doing something that takes 6 or 8 hours of computation like one of those Houdini simulations you were talking about not having error correcting can cause your system to crash in the middle of its work https://level1techs.com/video/memory-unleashed-threadripper-128gb-2933-ecc-tested The G. Skill is on the way any day now. But I still have some decisions to make about the M2 hard drives and the CPU cooler and the kind of case it will dictate. Basically I want to control the confusion by packing everything into an enclosed server rack (should be up on casters for ventilation and moving it around, like the external UPS, the routers and switches I want to experiment with for Cisco networking, another older server I own, and a hot swappable hard drive storage. So I'm looking for a 4u case and cooling that will fit that. Something like this stuff; I really want order and an easy to keep clean and dust free setup. But I'm outside my knowledge on this stuff...I'm asking all my server room friends for guidance on this...I think I'd actually prefer not to get into liquid cooling...just to avoid the maintenance.. And then there's the monitors...

Even with a major snowstorm last week this stuff got delivered really fast.

And it looks like at least initially, the NAVI won't be rivaling the 7nm Vega architecture for high powered computing and desktop workstations. NAVI appears to be focused in the near future on gaming, specifically targeting the future consoles. OTOH the new Ryzen X570 motherboards for Ryzen 3000 desktops will have PCIE gen 4.0 and The 8 core models can sell for around $199-$299 while the 12 core parts can go for $399 and finally, the 16 core parts can end up around $499. The reason we will be looking at such good prices is that unlike Threadripper CPUs which use a bigger PCB and four dies (based on EPYC layout), the Ryzen CPUs will only be featuring 2 dies and that saves up space and design costs. Also, the 6 core and 4 core parts may end up under the $150-$200 US bracket which would make them an ideal choice for budget users. https://wccftech.com/amd-ryzen-3000-cpus-x570-motherboards-and-radeon-navi-gpus-7nm-launch-rumor/ You can never be up to the moment very long in this technological race. At some point you just have to stop and say what do I need to do the job and then be happy with that.

My oh my, look what just arrived in the mail!

I just pulled the trigger on that G. Skill RAM. No regrets. Some guys blow $60k restoring some car. It's not all just dollars and cents; there's an aesthetic there. I'm 65. I'm going to be dead soon. Lots of my friends are kicking the bucket or wish they had kicked the bucket after a bad hip replacement surgery left them limping around in constant pain. Then there's the sorrow of not being 18 with a 16 year old girlfriend. But that's not me. I'm going to get this rig and do it right. It may very well be my last rig, my best toy ever. When I bought my new house my next rig was at the center of my ideas. It involved running three 4k monitors, 30"+ This is how it's done. A girlfriend is just going to take my house and bank account and pensions and leave me sleeping under a freeway overpass. I'll have to have a knife fight with some evil bums over who gets to sleep on the abandoned sofa. I feel sorry for Al though. That desk sucks. I studied the problem of a desk that can accommodate a 3 x 32" 4k monitor setup like this and came up with the ultimate solution; Restoration Hardware's Monastery Table, (AKA Partner Desk, Library Table), 4' x 10'. Even while I was sleeping in the parking lot waiting to take possession of my house, I was planning its purchase. It looked like it would be $6000 for the table/desk and $1000 for delivery. Then RH sent me an email on their spring sale. I nailed that table for $3750 and $250 delivery. This guy built one and it's nowhere near as nice as mine and he regrets not charging $11k USD for it. My only regret in this build is that it can't be the 2990WX; there's just too many problems with that thing. It's a mangled Epyc and the performance just doesn't match the extra cost. As for the RAM, Wendell at Level1 had this to say; https://forum.level1techs.com/t/picking-out-the-right-ram-for-2950x-shouldn-t-be-this-hard/132123/12 "See the video we did on 128gb ecc vs non ecc. Extensive benchmarking there. Some surprising results like 128gb of 2933 is faster for most, but not all, things vs 3200 32gb" Also have a look at this; $999 for a Pro Radeon that's nowhere near the GPU that $699 Radeon VII is. http://www.cgchannel.com/2019/02/review-amd-radeon-pro-wx-8200/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+cgchannel%2FnHpU+(CG+Channel+-+Entertainment+Production+Art)

I hear ya and I appreciate it but I'm all about risk avoidance.

No, I'm going with the exact X399 recommended RAM. Why? Because if they don't work properly try getting a refund on used RAM. $880 is a lot of money to risk on it. At this point I'll be going with the g. skill Flare x x399 2933MHz if the 3200 or 3600MHz don't show up. I've been closely following this matter on Level1 Forums and I'm going with Wendell's dictates on this one. There's been a lot of problems reported in this area with the 2950X and the 2990WX and RAM compatibility. I suspect it has something to do with the way the CPU is set up with those 4 x 4 core 7nm chiplets with their 14nm I/O die and their memory controller.

I haven't gone shopping for that yet; next on the agenda is RAM. 64 gigs in 4 sticks of 16GBs. The best I can find is the X399 specific G.Skill 2933MHZ Flare X. The 3200 MHZ and 3600MHZ are sold out.

Yeah but that's $344 Cdn. That's an extra $64 for another 100W. Remember, I'm buying all this stuff with worthless Canadian pesos as I slowly suffocate under a mountain of snow.

I just grabbed this; it had the best reviews of all the 1200W PSUs on Amazon and NewEgg and I found one last new one in Vancouver for $280.00 Cdn. It's got 6 four pin 75V VGA sockets on it so I could feasably run 3 x Radeon VII cards with it. I may get a second GPU. Not sure right now. Considering that other PSUs rated this highly were over $500 I figure this is a good price point. Auto Undervolt in the AMD Radeon Settings interface effectively deals with excessive fan noise.

Gimmie a call sometimes

There are risks on both sides. If 2000 series weren't dying all the time I'd have gone with two 2070's months ago. But here we are and I can no longer delay. Making a choice means having to come to terms with the shortcomings of both platforms. Space invaders or a noisy fan, RTX or 16GB of HBM2 VRAM and pro card drivers. Even now just finding the RAM I want at 3600MHz by G. Skill is impossible; X399 specific sticks are almost all out of stock save for the 2933 16GB sticks. Getting the PSU I want is hard too. I've got to get this thing built though. I've waited long enough.

Yessss... https://wccftech.com/amd-radeon-vii-graphics-card-radeon-pro-software-enterprise-support/ All the reviewers are dumping on it but that didn't stop it from being sold out on day one. At $699 I wonder why that was? kek.

Maybe you already know all this; I knew that the M2 was a great new form factor but I'm just realizing how insanely fast it is. https://www.pcworld.com/article/2899351/storage/everything-you-need-to-know-about-nvme.html There’s a reason why we still have SATA SSDs and NVMe SSDs. Knowing the potential of memory-based SSDs, it was clear that a new bus and protocol would eventually be needed. But the first SSDs were relatively slow, so it proved far more convenient to use the existing SATA storage infrastructure. Though the SATA bus has evolved to 16Gbps as of version 3.3, nearly all commercial implementations remain 6Gbps (roughly 550MBps after overhead). Even version 3.3 is far slower slower than what today’s SSD technology is capable of, especially in RAID configurations. For the next step, it was decided to leverage a much higher-bandwidth bus technology that was also already in place—PCI Express, or PCIe. PCIe is the underlying data transport layer for graphics and other add-in cards. As of gen 3.x, it offers multiple lanes (up to 16 in most PCs) that handle darn near 1GBps each (985MBps). PCIe is also the foundation for the Thunderbolt interface, which is starting to pay dividends with external graphics cards for gaming, as well as external NVMe storage, which is nearly as fast as internal NVMe. Intel’s refusal to let Thunderbolt die was a very good thing, as many users are starting to discover. Of course, PCIe storage predates NVMe by quite a few years. But previous solutions were hamstrung by older data transfer protocols such as SATA, SCSI, and AHCI, which were all developed when the hard drive was still the apex of storage technology. NVMe removes their constraints by offering low-latency commands, and multiple queues—up to 64K of them. The latter is particularly effective because data is written to SSDs in shotgun fashion, scattered about the chips and blocks, rather than contiguously in circles as on a hard drive. The NVMe standard has continued to evolve to the present version 1.31 with the addition of such features as the ability to use part of your computer’s system memory as a cache. We’ve already seen it with the supercheap Toshiba RC100 we recently reviewed, which forgoes that onboard DRAM cache that most NVMe drives use, but still performs well enough to give your system that NVMe kick (for everyday chores). This ASRock Fatality X399 motherboard has three slots for the M2. https://www.vuugo.com/asrock-motherboards-FATAL1TY-X399-PROFESSIONAL-GAMING.html - 2 x Ultra M.2 Sockets (M2_1 and M2_2), support M Key type 2242/2260/2280 M.2 SATA3 6.0 Gb/s module and M.2 PCI Express module up to Gen3 x4 (32 Gb/s)* - 1 x Ultra M.2 Socket (M2_3), supports M Key type 2230/2242/2260/2280 M.2 SATA3 6.0 Gb/s module and M.2 PCI Express module up to Gen3 x4 (32 Gb/s)*

It was possible to recreate node networks in ICE and then replicate them in Houdini and vice versa, more or less. Too bad the same isn't true for Substance Designer and Houdini.

Man this Corteks guy sure has a dour view of the future...

https://www.ebay.ca/itm/NEW-AMD-Radeon-VII-7nm-Gaming-GPU-DirectX-12-HBM2-16GB-W-3-FREE-GAMES/183674854944?hash=item2ac3dfde20:g:bYkAAOSwF8BcXG~i:rk:1:pf:0 Somebody flogging off their review copy? It's got the 3 games included...

You pushed me over the cliff ya son of a gun. I just pulled the trigger on that Radeon 7. I think one of the deciding factors is that it never was meant to be a gaming card. It's a cut down version of the Radeon Instinct, a scientific research card that sells for more than $10,000.00. It's got all the high end error correction stuff taken out which isn't necessary for artists so it won't be competing with the Instinct in that arena. I think this is going to be one hell of a lookdev card for sure at the very least. $1059.43 in Canadian pesos. Everybody else, even in the USA, was sold out except AMD themselves. Delivery in 3-4 business days. And that's with the 3 games for free; not my type of gaming but nonetheless the aftermarket producers don't seem to be offering that deal, only the FE. ONE PER CUSTOMER

That's yesterday, new, off Ebay

Very good question. We're going to have to research this. https://forums.evga.com/Nvidia-has-updated-the-Reference-Founders-2080ti39s-to-utilise-Samsung-GDDR6-m2879642.aspx https://forums.geforce.com/default/topic/1079421/geforce-rtx-20-series/rtx-2080ti-new-batch-gets-samsung-memory-/ Just watched that Linus video review. He seems to think the Radeon VII might be the ideal card for content creators..

With GeForce RTX 2080 Founders Edition subjected to similar methodology, AMD's flagship ended up a full 10 dB(A) higher. What AMD asks of its cooler in order to match Nvidia's performance is not acceptable at the same price point. Tom's Hardware

I was just reading WCCF Tech and hit refresh and this popped up fresh off the press https://wccftech.com/amd-radeon-vii-worlds-first-gaming-7nm-graphics-card-review-roundup/ I'm not really sold on this; I'm thinking the RTX2070 if you can find it with Samsung VRAM instead of Micron is probably a better deal, especially if you buy two of them. More troubling, we believe, is Radeon VII’s acoustic situation. Following in Nvidia’s footsteps, AMD retired its blower-style cooler in favor of one with axial fans. But rather than creating a quieter thermal solution able to keep Vega 20 cooler, Radeon VII is easily just as loud as the reference Radeon RX Vega 64 due to a fan curve that ramps up to 2,900 RPM under load. We approached AMD about Radeon VII’s noise because, frankly, it’s disappointing. The company explained that the card’s shipping configuration is tuned for enthusiasts, and that it’s working on other options that’d conceivably trade performance for better acoustics.