J7SC_Orion

...sometimes Canada stands in for Russia (and elsewhere...) in James Bond titles...Mt. Asgard, you can visit w/ FS2020 ?

..needless to add that I'm really looking forward to sink my teeth into this (don't have any background w/ AM4, usually just do AMD TR and Intel HEDT). The 3950X does seem to have a strong IMC +IF though...and I used one of my 'early' 32 GB GSkill 3866 kits (which is known to go up to 4133 before voltage increase in other builds). Problem is time - hyper-busy at work...wish I could just dive into it now.. ...all stock re. voltages. Haven't done CPU oc yet, been focused on IF / RAM

@ENTERPRISE ...great news about the upcoming marketplace, with special thanks also to @axipher & @Avacado ...cross-border transactions soon then... ..."From Russia Canada with Love"

...quick update on 'sealing' push/pull with electrical tape: I'm just putting together a X570 Asus Crosshair Hero VIII wifi and a 3950X...got a BeQuiet PureLoop for a very low price (and it's even refillable !)...while this AIO is a great performer as is, I took the BeQuiet fans off and mounted six Artci P12 pwm pst in push / pull. Even at full song, you can hardly hear them and manages to keep a 3950X in repeated Cinebench23 in the low seventies C. This unit will get a full custom loop as soon as the 3090 Strix block and backplate is settled on, but for now this is a more-than-decent solution. Pics below of the testbench for that, not final...note the electrical tape on top of the rad...still have to do some trimming, plus the sides...

^^That ! For a variety of technical reasons, I usually go with dual D5s per loop anyhow, but there's nothing stopping someone to add a D5 to one that comes with the distro. BTW, there are several 'custom' distro plates by ThermalTake for the Core P5 as well as Bykski and Barrow - or the whole Core P5 '''case''' is a distro...first shown several years ago, now finally available...must be pretty heavy though...

Nice ! I have a work-play TR4 / X399 MSI Creation setup that does all the heavy lifting...most reliable and powerful thing I built ever ..tx - was accidental more or less... tried out the 3950X and saw one core at over 4.7 GHz after 1st boot-up, a few others not far behind. For now I'm using a BeQuiet 360 AIO (but with six Arctic P12s in push-pull) and highest temp w/ 24 C ambient was 73 C in stress-runs of Cinebench 23....just got to get used to that idle voltage ?

...started to have irregular boot issues with my 6yr old X99...no time to fool around on a hybrid productivity / gaming system right now...so, I got this, at pretty decent prices actually. 3950X instead of 5950X, but at least it was actually there for pickup (1 of 2 left) instead of 'out of stock'. Custom w-loop comes later...

Tx ...I hope so...I did that with my 200GB cache but every time I go from one system to the other (never concurrently), there's all kinds Microsoft 'validation' going on...syncing both at the same time would be perfect.

....great ? .I get to download 'massive' twice, for two builds ?

...check out DerBauer vids on YT on this (multiple O11 builds, rad configs)...after all, his name is on the O11, too ? more > here

One quick tip when talking about positive pressure, push pull etc with PC rad fans in particular. More often than not, I noticed a tiny gap between the fan 'body' and the rad, often because of the vibration taps, sometimes because of the fan body design in general. With my GentleTyphoons in push/pull, I actually use electrical tape all the way around the edge of the 360 rads to seal that gap between the rad and fans...not scientific, but I think I picked up a 1 C+ drop, and also slightly better noise readings ?

Thanks ? ...good to know. I'm considering the Bykski for my Strix

...it's a long weekend here...lounging in front of a 55 inch TV with a big cup of java and pushing the 3090 a bit in FS2020. Enjoy the real-weather NY scenery:

" As it happens I still have some of those Gentle Typhoons hanging around never used lol. Wish they were PWM but generally speaking a decent DC control is fine if through the motherboard or external controller." ...ohh, but there's a solution: https://www.overclockers.com/high-speed-gentle-typhoon-pwm-mod/

...re. your question as to whether push+pull makes a difference, 'yes', but it depends to some extent on the rad setup (thickness, fin density) and of course the fans themselves. I build my first quad-SLI back in '12 and used a custom 1080x60 rad (equivalent to MoRa360) with 18x static pressure fans, 9 at the front and 9 at the rear...the difference on that particular setup was about minus 5 -6 C for the CPU under load but more pronounced when loading the Quad-SLI with benches. ...same with the 2080 Ti SLI-CFR setup, it uses 3x 360/60 in push+pull with GentleTyphoons (and one w/ ML 120s in push pull)...static pressure isn't a problem and the resulting 'noise' is largely mitigated by baffles I added (which also help airflow, btw) as those rads sit 'on their side'. Key is to have 'matched fans' running on the same controller - you don't want the fans to hinder each other ...with the Arctic P12 PST I'm just getting into, installing them on one side dropped temps by over 4 C compared to the ML 120s - and the P12s are amazingly quiet even at speed...adding push/pull to those tomorrow in the new build and will update here ? ...in the meantime, check this out:

Amazon and eBay for the 3090 Strix OC below as of today...I could more than double my money in two weeks...nutty times we live in. Point being though, keep looking for a good deal, there are some to be had behind all the scalper noise...

...some excellent points ! However I prefer not to mix anything into the final fluid combo before leak-testing. As I build both servers / workstations as well as fun machines, I have a setup with an external Molex that powers all the pumps etc for a new w-cooling build w/o having to turn the main mobo on. I let that run for several hours, power and cool down, than run for another few minutes - be that for a single loop or dual loops. More on fans below... ...I have a collection of the Noctuas - but that colour ! I use them for workstations that go into 'boxes', but for open builds, no thank you. I got so desperate on a personal TT Core P5 build that I actually spray-painted some Noctuas black (still work, btw). ...for benchers and heavy duty, I use the tried and true GentleTyphoon 4K rpm models, each w/ 183 CFM...they are for example on my GPU loop for dual 2080 Tis and they just keep everything below the first speed-bin loss...so far, I haven't been able to 'heat-soak' that GPU loop with 760 W in 2 years, and not for lack of trying ! ..we also have a big pile of repatriated Sunon 5K rpm server fans, but while they certainly cool very well, they just can't be used if you sit within earshot, whereby I find the GentleTyphoons 'somewhat bearable' ? ...GT 4K on the left, Sunon 5K on the right: ...lately however, I have been getting into Arctic P12 pwm PST fans - and I'm genuinely shocked at the great price-temp_noise performance ratio, especially if you buy the 5 value pack. Obviously I don't know yet how reliable they are over the longer term, but they sure beat the cr_p out of for example Corsair ML 120s, both in terms of noise at full tilt, and temps w/ all other factors held equal. I just put 5 of those into 'The Thing' (you know as the Strix 3090 OC testbench) - 3 for the CPU, and 2 for the GPU backplate (double-sided GDDR6X = toasty). When I watercool the Strix, I'll be using those. The question is whether they last as long as for example the GentleTyphoons (some of those are over 8 years old, & in continuous use...).

..lately, I only do 'open' builds (unless workstation / server use by others). The TT Core P5 started this trend with me in late '18 (dual loops, dual 2080 Ti, total of 4x D5s, 5x 360/60 rads) and while not super easy to stuff everything in/on there, it was a ton of fun. With a regular case, that would have meant problems galore. Two more builds since then for my home-office, all open ?

...overall, top end GPUs are moving towards explicit mGPU anyway (Intel, AMD, NVidia) but lack of SLI support now at a time most folk can't even get a top card is 'bad' timing

AFAIK, GDDR6X still has only one supplier (= pricey) and also needs additional quality control step$

...yeah, may be it's just a love for some degree of symmetry, but my fav builds to make are HEDT (Quad Channel RAM) and dual GPUs, like below...been doing SLI since '12 ?

...yeah, but they're doing their shareholders a lot of favours with that. For the regular consumer though: ? ...also, I don't believe that in NVidia's financials, gaming and mining card revenue are separated (and some folks do both anyhow), while enterprise level GPU is a separate column

There are too many factors interlocking right now to even try to forecast mining in particular and GPU prices in general...including China's introduction of a digital currency, 'casino speculators' galore - but also a real attempt by folks to hedge a bit against future inflation, given latest money supply growth figures. Then there's that trade war with a 25% tariff...the higher the base price, the higher this extra 'tax'. Also, it looks like NVIdia (and soon AMD ?) have successfully established an 'ueber-class' of GPUs which they can't produce fast enough to meet demand...they would be foolish from a profit-maximizing POV not to exploit it, as much as I don't personally like it Add several confirmed news stories about hundreds of millions of dollars of Ampere being bought up right at the factory gates where they are manufactured and the plot thickens...that's before additional demand via Covid-19 factors impacting both the home and workstation markets, or the higher input prices for components of a GPU (and other electronics <> note the shortages of some bits these days) Speaking of the workstation market, and this may sound a bit lunatic, for what you are getting in terms of performance, even 3090s are a great deal from the POV of a productivity purchase. I use my 2080 TIs and the 3090 for both work and play and know the prices of a workstation Quadro RTX. I wanted to add a 16 GB min VRAM GPU as I'm moving completely to 4K and was set to find a 6900 XT which I really like, but software support and also just the 4K performance drop-off steered me towards to 3090 24 GB VRAM, so a partial business purchase. A final point which bugs me a bit is that the OC community (to which I fervently belong, including prior LN2 XOC) should look at itself as well in this pricing and availability context. The community made up of both responsible and less responsible folks. I know of several people who just order and/or pick up a 3090 air-cooled model, slap the KPE XOC 1000 W bios on it with all safeties removed for bench-marking - then return it to the store a week or so later if it wasn't a 'golden chip' in their opinion. The store then resells it as open-box (presumably with the old bios flashed back). Others go through various hard mods (shunting etc) and that's fine, I've done hardmods myself - but it is a risk. .I would never try to take the mods off if the card dies and return it to the vendor, but others do...vendors are not stupid and often catch the (former) mods, but in other cases they just replace the card. Ultimately, the RMA cost is added to next year's pricing by the vendors, again as a percentage, so the higher the price of the ueber models, the higher the RMA 'premium' will be. If mining slows down sharply, I do expect GPU prices to flatten if not sink a bit, but the other factors at play will continue to beset the markets for a while, IMO...

...the sooner the better ! Yummy:

...once I though I was wrong, but I was mistaken ? We are living in amazing times - I bet people will even buy the 2060 and 1050 ti at inflated prices...