J7SC_Orion

What GPU are you running ? I ask because FS 2020 was a DLSS3 Frame Insertion launch title for the RTX 4090 (and presumably other RTX 4k). With the 4090 and DLSS3 and Frame Insertion, I picked up an additional 50 fps +- over my 3090 setup with the same planes / same scenes.

...I focused primarily on my fav existing aircraft and known scenery for now to test out the promised performance gains and fixes...DX12 was already more or less playable for me before but clearly things have much improved under the hood in addition to the DLSS3 Frame Insertion goodies and the steps they took to control latency. Marketing bumpf is one thing, actually testing it out on my setup was the kicker...

...I wish they would apply that to AM5, and also improve the memory controller to get much faster and catch up with Intel's 13th gen...I might even buy an AM5 then (or earlier if/when they release a 7950X3D ). Speaking of memory controllers, Level1Techs had an interesting observation re. the new Epyc Genoa (up to 96C /192 T, DDR5, based on the Zen 4 chiplet) having a better memory sub-system - and I'm not talking about sheer bandwidth of the 12 channels per below. Possibly some new microcode that hasn't made it to the Ryzen 7K yet...

Folks, if you haven't already done so, install FS 2020 SU 11 (November 11). It's a massive ~ 24 GB patch, but worth it !! RTX 4090 in 4K HDR ultra everything and 400/400 detail hits almost 120 fps (4K OLED limit) with DLSS Quality and 'Frame Insertion' !

I can feel your regret (if that's the right expression) for having to pass on that 4090, but there's that expected 4090 Ti for you soon ...my Giga -G-OC can manage 3 GHz on 1.05 V / 450W, but once I kick voltages to 1.1 V and PL to 133% (up to 600 W on stock vbios), it can reach far higher - below is just a heat-cycling run (from '3DM Speedway above) at 25 C room temp after installing the water-block, and it is not yet top speed...but the flatness of the clock lines is nice and courtesy of the water-block. It really needed a water-block for a variety of reasons, including some questionable factory thermal pad placement, factory TIM application and such... But as you know, benching for me is not as serious an undertaking as it once was. The real reason I got the 4090 is for the handful of games and sims I have time for (and run at 450W max). Best example is FS 2020 and its SU 11 update from a few days ago. With everything-maxed, 4K Ultra / DLSS Quality and the RTX4K-specific 'frame insertion', I picked up around 50 extra fps...even in difficult scenes (flying low over complex metro areas), I have seen 110+ fps. This near-maxes the C1 OLED, which is what I wanted as a chain is only as strong as its weakest link. You might say the AM4/5950X is the weak link now, but with the latest NVidia driver and feature magic, it doesn't matter anymore. So now I can take a look what to upgrade to re. mobo and CPU with peace and quiet when the time comes while enjoying top graphics already on the OLED.

...haven't subbed some of it yet; just doing heat-cycling at 25 C ambient in the room after finishing the water-block install today (per below, tubing and wiring needs to be cleaned up)...may be next week I turn the heat off for a bit and make everybody here freeze at night Also looking forward to your new GPU's results !

...several hi-po PSUs (ie. some Superflower / Leadtek) have 9x or so...and below is a pic of the back of the PX 1300 I use. I also just finished the water-block conversion of the 4090 and installed the new cable today - system still needs a lot of clean-up re. tubing and wiring...just doing heat-cycling right now...

The more competition, the better ! Gamers Nexus also had a segment on this, per timestamp below, the new company was founded by a former AMD VP - Mr. Leatherjacket is probably not too thrilled...

...have been prepping the RTX 4090 for water-cooling as all the parts and peripherals have arrived. In the meantime, I've been tuning the 5950X as I discovered a few more secret sauce ingredients . RAM is now at > 3800 and 14-13-13-13. @Bastiaan_NL, I've got a screenshot of one core actually peaking at > 5.7 GHz (on water, stock CPU voltages, same as below), though that was more or less accidentally...I am still debating whether to keep this as the primary play setup or move to a different platform...I wish AM5 would have a better DDR5 memory controller (or get on with it and release a 7950X3D).

...ahem, obviously, they have to match on the cable end and card side connector (it's a male - female thing...). What I am trying to underscore is that there were different connector styles (some with / without) recess that were harder to fully seat and also left potential gaps. My 4090's connectors worked perfectly so far and are in pristine shape (per pic) even with hard benching and spikes > 600 W because they weren't twisted, and were the right type from the get-go (no recess or gaps). In addition, one set of the new custom cables is already here anyway (again meeting 'option 2', as was the original). So there you have it, my last comment on (various!) 'cablegate' threads at this forum

I am talking about the gap where the sideband cables dock...on some cards, it was/is recessed and thus potentially leaves 'room for error' (the 'before' pic in the drawing) whereby option 2 having "only a shroud extending beyond the plug body". As you can tell from my pic above, it is flush.

...it looks like my 4090 already has that (option 2 per link above) . FYI, the first 12VHPWR replacement cable has arrived now, the 2nd one (a freebie from Seasonic) is on the way. In any event, I just removed my 4-into-1 dongle connector for the first time since I got the card almost three weeks ago. I had done a slew of early benchies (up to 575 W actual per HWInfo) at 25 C+ ambient....Looks pristine to me, though I had made sure a.) no sharp twisting / bending and b.) fully insert and 'seat' the connector during the original install.

...some more 600 W management tools

...one of two custom cables for RTX 4090 / Seasonic P.Plat. PX 1300, the other (black) is a free upgrade by Seasonic but hasn't arrived yet.

I am not defending Intel's (and NVidia's) decision making...fyi, below is a 7 month old vid from Linus Tech that lays things out for those who are interested. That said, if you want to go beyond 450 W / 520 W (ie. spec for 600 W) as a vendor selling to the public, you can't rely on / legally use the old PCIe 8 pins - unless you use 4 of them - per certification process ! It is another discussion if we actually need 600 W GPUs... For the record, I am thrilled with the RTX 4090, and while its 4-into-1 dongle can have issues if quality control was out to lunch or the user did not seat it properly (two separate issues), I was careful and have no problems with it (apart from dedicated 12VHPWR cables arriving this week for my Seasonic Platinum PX 1300W). Also keep in mind that the 4-into-1 dongle actually runs with up to 4x PCIe 8 pins, so the PCIe 8 pins are still 'in play', anyway. ATX 3.0 PSUs will be 'fully native 12 VHPWR' at both ends.

Ouch !

Something IV-related actually happened to me last month...per earlier comment, I had a back operation, and in prep for that I drove up three days earlier for their lab to take blood. The lab is across the hall from the surgery area on the same floor - and both up an a mezzanine with open view to the main entrance / check-in area. As I was already hooked up to the IV, they realized that they couldn't find my blood test (from across the hallway, three days earlier...). So I had to '''march'' down to the lab again while already dressed in that weird gown (open back) along said mezzanine with the IV on a stand with wheels trucking along. To top it all off, this time, the nurse taking the blood was an absolute beginner, making multiple attempts - painful blue-black blotch on my arm for two weeks, unlike the first time in that lab.

I agree that sometimes, Intel tries to push certain 'innovations' that aren't useful, nor accepted by the marketplace. Still, Intel's specs basically rule modern motherboard design, and there is also the increasing legislative push that may affect power consumption (re. emissions). On the 12VHPWR native connector, there are a lot of new PSUs coming out with that via ATX 3.0 spec.

...just like more car systems are being switched to 48V, that would make sense. But in any case, folks seem to forget that the 12VHPWR connector is an Intel spec (below, incl. 600W). The real issue is that it wasn't introduced very well by NVidia, and apart from some quality problems with the (temporary) 4-into-1 dongle, there also seem to be a lot of cases where the connector was not plugged in all the way on the card side and lock in with the hooks, leading to aforementioned issues. That could be because of quality issues (such as production tolerances) or other reasons, such as user error out of an overabundance of caution, ironically.

...side-mounted connector would have been easier in some cases, but the 4090s air-coolers (unlike PCB) are way too big lengthwise - even w/o side-mounted connector, they don't even fit in many cases unless water-cooled (much shorter). Anyway, My 4-into-1 dongle is securely inserted and attached, and I am also getting 2 sets of 12VHPWR native cables delivered this week.

... This > is what I was referring to above re. not fully inserted connector. In any a case, there's also this

...latest 'suggestion' in several forums relates to the connector (even a native 12VHPWR one) not being fully inserted.

...the block comes with pads and a backplate (which I will modify w/ an extra heatsink). That said, for the VRAM and parts of the VRM, I will use TG-10 thermal putty again which easily conforms to any available space. I also have a selection of various sizes of Thermalright Odyssey pads from earlier Raven builds left over (see left pic).

...a Bykski water block for the RTX 4090 Gigabyte-G-OC, along with two packs of Gelid GC Extreme for the GPU die (VRAM gets thermal putty again, phases some Thermalright pads). It's funny how small the PCBs are on the 4090s once you see it w/o those giant, heavy air coolers. I really don't want to run a card that ranges from 450W to 600W on air.

I had a back operation last month - went well, and a related biopsy came back negative (the good kind) - but I still don't like hospitals much (no offense to anyone). I had a great team for surgery, but the lab next door on the same floor had lost my blood test they took three days earlier, and general 'online/paper work' took longer than the actual operation. All is well that ends well, I suppose.