J7SC_Orion

Those valve fittings are a smart idea with a mobo monoblock ! I remember trying to drain a loop with "just" two GPU card blocks and mobo_vrm w-cooled setup for an upgrade, and that was some theater (in the Greek tragic sense). Since then, I do discreet loops for GPU(s) and CPU/VRM anyway, but valve fittings would do the trick even better. Another quick question: Are those hoses / tubes in your pic EK ZMT, or other ? I would like to use EK ZMT for an upcoming workstation build, but also read that EK ZMT only plays nice with EK fittings (not sure if that is true).

Nice. I also like placement for the D5 Next with quick readout...and I gather you did not have to use the new O-rings (good to have some spares, though) ? Another quick question: Is that some sort of valve (green circle) ?

yeah, but I'm not cutting my ear off !

...yeah, definitely still very much a niche market for SLI/NVLink, as is 4k (according to Steam user data). Perhaps with Hopper onward, it *might* change for the aforementioned reasons. As to running remastered Crysis, possibly, if it uses the Cryengine updates in Neon Noir

...for now they'll probably keep it more in the shadows, with support still spotty depending on app...but per above posts, this is probably more for SW developers for 'some future-gen GPU architecture (such as Hopper)' when mGPUs / multi-chiplet designs come out. Still, some apps work great already - and yes, "it can play Crysis" EDIT: Not so much about CFR vs AFR/2, but about SLI/NVLink in general vs a single card...: Here's an interesting YouTube vid (by BFG / Benchmarks For Gamers) about 2x 2080 Ti in some modern game titles. Looking at 1% low FPS, when gaming with 4K especially (rather than 1080p), the results are pretty impressive

Well, some select 4K results now for Crytek's Neon Noir... First, in case you're wondering why NVlink_SLI 'CFR' is also called checkerboard ...not everything works as it should if you 'force' CFR...here is some Firefox browser fun ...but Crytek's Neon Noir bench works very well with CFR ! Not only better scores but apparently also lower and more consistent frame times. 4K results. Please note that there was no overclock (GPU or VRAM) on the GPUs in order to keep it all compatible between runs and also between resolutions. Typically, full-tilt OC with this bench and setup adds about 10% -15% to the scores. Also, this might still be the older Crytek Neon Noir engine (I thought I updated it, but every time I log in, it wants to update again - still, the same engine was used for all the runs below). Settings: 4K, Ultra, Fullscreen, stock clocks 7073 - Forcing CFR - 6326 - Forcing AFR2 - 4282 - Forcing AFR - (AFR is default dual card setting) 4733 - Single/Auto-select GPU -

...I saw that earlier...but 1.194v @ 7.7 GHz ? I know liquid helium is great for oc'ing, but... ...most likely, it had some power-saving feature kick in. In any case, great achievement by Elmor ! Looks like the ol' 14nm +++++++ still has some life in it

...ahh, so the o-ring is the culprit !? Depending on cost, may be get both sets (55z4, 55x5), also for future reference ? BTW, here's another 'gunk' item that can be helpful, for example to (thinly) coat older gaskets and o-rings with. It's non curing (unlike the Seal-all above) which is good for subsequent disassembly. This is the polyurethane version; there's also a silicone-based version. I have this before in 'universal VGA blocks' after taking them apart after noticing a small leak due to a worn o-ring seal (...again, if used, should be thinly applied, and double check list of materials for compatibility)

Bummer ! But at least you noticed it early enough. Loop pressure should not be able to do that within normal parameters. Can you actually tell exactly where the 'leak' is coming from on the reservoir (i.e. is the area where you see it also the source or just the collection point due to gravity etc) ? Or could it possibly even be condensation, or does it match your loop liquid ? I ask because I have used the stuff below on leaking cooling blocks for a PC to great effect, as well as fixing a huge fuel line for the carb in a car I had at the time (the fix lasted for at least another 200k km before I sold that beast).

...WC GPU w/ D5 Next - nice... One thing on my mind is the 400W power rating of the A100 'big' Ampere at ~ 1.4GHz. Granted, 54 billion transistors, 40 GB of HBM2 and all that stuff which likely won't make it into the consumer 3080 Ti or Titan versions. Still, with 7nm designs concentrating hot-spots, it will be crucial to find a really well-designed GPU water-block.

...so that's the monster under the bed kids are scared off ?! Found some more detailed info on the (current) top 'industrial' Ampere, the A100...For 3080 Ti (and Titan Ampere), take out HBM2 40GB for 24GB / 12GB (?) GDDR6, take out a few tensor cores and add rt cores - and voila. Real question is what final configs they'll launch, also with an eye on big-Navi (also due in Q3). All that said, by the time custom PCB cards are out and enough supply is available, we might be into Christmas holidays / early 2021 ? Hopefully, it will be a better release than RTX Turing was, and no 'test escapes' this time...

...well, a project for a rainy weekend ;-) In any case, the MSI X399 Creation, with which the Aoero 4-slot M.2 Xpander card came, obviously supports it, at least if I'm willing to give up some lanes of the secondary GPU. On the other hand, the Asus X99-E WS (pic / brief description below) comes not only with 4x PCIe 3.0 x16, but also with native M.2 NVME onboard...besides, unlike the MSI X399 Creation build (and the X79-E WS also in use), it currently doesn't have SLI GPUs installed anyway. On a related note, before I take the plunge, I should confirm that new 2TB M.2 NVME PCIe 4.0 drives will work in the (PCIe 3.0) MSI Aero Xpander, albeit with the drives in PCIe 3.0 mode only

Thanks guys...I have one MSI X399 Pro Carbon as my sole X399 "reserve", plus my soho office has both an Asus X99-E-WS and Asus X79-E WS humming away at daily tasks. PLX is onboard w/ at least the Asus E WS mobos

Yeah, I would prefer to keep my two 2080 Tis at 16x each, though theoretically, with the TR X399, there should be enough lanes. In any case, if that doesn't work right (ie performance loss on GPU NVLink), I just plug the Aero Xpander 4x M.2 into another MSI X399 mobo I have that carries a single GPU - or perhaps even in an older Asus X99-E-WS (Intel), if I can get the drivers going. That should also actually work (w/ the right drivers), given that this is 'just another' PCIe spec card...!?

Thanks much ! I should also check the mobo manual for the X399 Creation re. 'the 3x regular' M.2 slots on the mobo and which ones connect to the CPU vs chipset

A bit more on mGPU CFR, AFR/2... - Rumour has it (rumour = add salt shaker!) that NVidia, very quietly, added CFR for developers working on 'Hopper' (architecture after Ampere). With GPU dies still being huge even after node shrink to 7nm, mGPU is expected to gain some wind in its sails as vendors *might* have to focus on multi-die single socket GPUs (a la AMD Ryzen CPU pattern) - Below are 2 x 2 benches (Valley, Neon Noir) comparing CFR and AFR NVLink modes. For now, just at 1080p, no GPU OC etc which is also held back a bit by the TR 2950X at that resolution (unlike 4K, but that monitor is in use on another test-bench). Then again, the same handicap applied to all tests, and other factors such as ambient temps were the same. Still, I look forward to add full-OC, 4K results. - CFR seems to work in Valley, Heaven, Superposition and Cryengine Neon Noir..and work very well. I noticed not only slightly higher scores but also a bit lower frame times Valley 1080 AFR Valley 1080 CFR Neon Noire 1080 (ultra) AFR2 (better than AFR for this bench) Neon Noire 1080 (ultra) CFR ...and for 1080p AFR 'all stock' vs 4k full-OC AFR (incl. GPUs + CPU) with the same quality settings (but slightly older driver)...

Below... 1.) Updated DLSS performance (3DM Port Royal as before, but added DLSS feature test) 2.) Thinking of installing the MSI Aero 4x M.2 add-in card that came with the X399 Creation mobo for extra storage (7x M.2 total)...just need to figure out which slot (given dual GPUs) and whether to trust the PCIe riser cable that came included with the TT Core P5.

IMO, nuclear power technology has matured with newer and very different systems that either use waste from the prior (current) nuclear plants, and/or do not produce the same type and level of byproducts and waste. They are also much safer, with emergency shutdowns based on natural physical laws, rather than relying on a (possible submerged or damaged) diesel generator... One problem seems to be that a lot of public and private sector utilities placed humongous financial bets (50 year bonds etc) on the original 'dirty' tech which stemmed from earlier applications (such as nuclear subs) where space was at a premium - and in the process generated a horrible environmental reputation in the public, not to mention Fukushima Daiichi, Chernobyl, Three Mile island... If they can solve the hydrogen storage and transportation issues, I do believe hydrogen could be a game changer, not least as a lot of the 'globally installed tech base' could easily be converted (ie most types of internal combustion engines). This is akin to the battery technology race with electric mobility. Time will tell, though with the current 'cov-virus' leading to massive public-sector debt and reduced economic activity, energy-related R+D and implementation budgets could get quite dented

Yeah, the RTX launch was a bit odd. That includes the closely-spaced release of the Titan RTX with only marginal gains for twice the costs compared to the 2080 Ti (unless the extra VRAM was crucial to given apps)...perhaps all rushed because of the mining collapse ? I'm taking my time re. a potential 3080 Ti upgrade as I want to know more about AMD's upcoming flagship, and even NVidia's 'Hopper'. Add to that the much stronger performance and applicability of DLSS 2.0 even for current RTX cards... That said, on the 'mini-enterprise' level, NVidia's Ampere DGX-3 'mGPU' actually looks like a deal for a given set of companies...8x A100 GPUs, 600gb/s NVlink bridges (non-SLI), 2x 64c/128t AMD CPUs and on and on...about half price of the current version but with much higher performance...since NVidia is not in the charity business, they must be expecting some stiff competition in that lucrative space from Intel and/or AMD and their next-gens (and are perhaps even pulling Hopper ahead for that). All in all, this should mean a whole busload of architectural trickle-down tech for consumer and prosumer cards

Tx...I also wonder what the specs of the 3080 Ti will be, including on power consumption....not least as the just-announced Ampere A100 7nm workstation card at 6192 Cudas / 40GB HBM2 is rated at 400W @ ~ 1.41 Ghz, according to AnandTech (repeated below) - and that apparently is still not the full-fat die GA100, which has not yet been revealed. 3080 Ti is unlikely to have the 6192 Cuda and respective Tensor cores (though it is all just guessing at this stage) and likely will come with a smaller GDDR6 VRAM set. But still - 400W @ ~ 1.41 for a 7nm is a lot, though it has 54 billion transistors...May be I should stock up on SuperFlower's LEADEX Platinum 2000W PSUs, HeHe

I know what you mean...I have been thinking about upgrading from the 2950X / X399 / dual 2080 Tis but there really is no need, at least not yet. The 2950X is a good sample (low-v, high oc, strong IMC) and I use 4K monitors, so GPUs carry a bit more weight in upgrade decisions. Also, I need HEDT work-wise. That superb 3950X / X570 Aqua should last you for quite a while; still, it is good to have an upgrade path if / when you want to exercise that option. BTW, how do you like that EVGA 1600 Supernova P2 ?

...cold and rainy Saturday + some Covd restrictions make for some good benching weather I re-ran my favourite (PortRoyal) and also the DLSS feature test for the first time (which is basically 2x Port Royal in a row, with / without DLSS). That makes it a nice long bench to test cooling system performance...all those pumps and rads did their job; GPU loop is not heat soaking and temps remained fairly flat at under 30 c / full load, which makes a big difference with 2080 Tis

Nice ! When I checked the pricing for the Aquacomputer D5 Next, I started to feel a bit guilty, but then checked other D5 pump and water-cooling gear prices - wow, what happened (rhetorical question) ? The good part about D5 is that they last a long time, but since we bought a bunch back in 2013 -2014 (same for XSPC RX 360 rads, fittings, GentleTyphoon and other hi-po fans), I haven't really kept up on pricing...just clean and recycle the old stuff . That said, if/when more D5s are needed, it's going to be the Aquacomputer D5 Next with embedded flow-meter. Looking forward to your full build pics - those Asrock X570 Aqua full-block mobos are cool(ed) and gorgeous (not to mention ready for the next-gen Ryzen)

Dual loops ? What is EHW coming to ? Elsewhere @ EHW, there was a bit of coverage of why it can make sense to have dual D5 pumps in series per loop under certain circumstances. As context, this includes fail-over (in particular when dealing with workstations and servers), as well as usually higher and consistent flow, and more resilience to air bubbles - especially if there are multiple rads and CPU+ GPU blocks. That said, D5s rarely break (unless they ran dry for extended periods), and dual D5 pumps are certainly not a necessity even for a high-end enthusiast build project. Still, I prefer build all HEDT (work and play) systems with dual D5s per loop. But what about separate loops for CPU (w/ perhaps VRM), and GPUs ? Why does that even make sense, especially if a given loop layout already has dual pumps ? From a cooling temp point of view, it really isn't necessary, and it also costs a bit more (extra reservoir, fittings, tubing, pumps...) But it can have advantages: - First, when building a complex water-cooled system with a single loop serving 4+ rads, a CPU block and dual (or more) GPU blocks, there will invariably be a lot of twists and turns via tubing and/or angled fittings. Splitting the loop into two discreet ones is easier to lay out (depending on the case of course) and avoid tighter turns such as 90 degree ones, and it also reduces the number of overall impedances per loop. - Second, loop maintenance becomes a bit easier (ie. draining, air bubble bleeding) - Third, changing, say, the CPU and/or mobo but keeping the GPU(s) later on becomes easier...ditto for changing the GPU(s) but keeping the rest - Fourth, it is fun ! - Fifth, I like anchovies on my pizza, so there... apparently, Iamjanco likes anchovies, too: ​ Do Orcas eat anchovies ? ​

...Tx. And in the interest of public safety, I also already warned innocent bystanders of potentially contracting pumpritis and fanritis after prolonged exposure <> here & below