tps3443

The EKWB 5090FE block is working wonders. So overall build quality is nothing to really write home about. It’s just a typical waterblock. But the performance is a different story. Water deltas or the GPU temp over water temp so far and what I’m seeing is really really great. 430-450 watts=7 or 8c delta 600+ watts=10.5c delta. 800+ watts=13-15c delta It has to be because of the Liquid Metal. I don’t really know. I used a heat gun and melted my thermal pads while on the block and then pressed the 5090’s PCB on to the block, this provided some good contact as well. Even drawing just under 900 watts of power, and it’ll hold a 15c delta over water. So with 34c water temp my 5090 pulling almost 900 watts and my GPU is running 48-49c (This is nuts). If I had 15c water it would be running 32-34c@800+ watts. This thing puts a block like the 3090KP HC block to shame. But then again, I feel like an idiot for never putting Liquid Metal on that GPU, I always used paste. I feel like it’s magical though. So it’s either the best waterblock I have ever tested, or Liquid Metal is magical lol. Either way, it’s performing really really really good. I’m not using a chiller anymore so I get to experience the feeling of warm water temps just like everyone else does. 😁 and with a shunted 5090FE now in the loop and a Xeon w3175X, and my Dominus VRM’s water cooled as well, I get exactly that (Some bath water) I will say though, I’m not running push pull fans which would help wonders for my setup. Fans are expensive. And it’s a lot cheaper to buy (8) P28’s than (16) P28’s. But now I’m feeling it and I think it’s time to add push/pull. This seems extremely cheap and dirty for AMD to do to save a few pennies. I’d want an original one. Luckily it’s easy to spot the difference in a used one, maybe not so easy on a new one though.

It’s pretty wild that Shunt modded water cooled 5090FE is doubling a RTX5080FE which is absolutely crazy. My RTX 4090 had nothing meaningful left in it, so nice to see the 5090’s definitely loves a shunt mod and watercooling.

My 2x8 to 12v-2x6 for my Lian Li Edge Gold 1200 in my PSU box in my closet. Maybe I should use it instead of the 12v-2x6 to 12v-2x6. 🔥

#29 HOF 3Dmark with my Founders Edition. 😁 NVIDIA GeForce RTX 5090 video card benchmark result - Intel Xeon W-3175X Processor,ASUSTeK COMPUTER INC. ROG DOMINUS EXTREME

Look at this thing man. Crazy. The silicon is so FRESH. https://youtube.com/shorts/POE4RDU_iL8?si=O_XzKLDZYpuLzE0Y @electrosoft https://youtu.be/40wojZVVGBk?si=YKkf-HSzPXFAh3_H

I wish I could bench it, I’m still working 🫠 but I did a couple runs and I’m already surpassing a RTX PRO 6000. 😃 I really don’t think it’s even that much faster. So remember RTX PRO 6000 card has a +200Mhz higher boost clock, and a +25 watt power advantage. So in those titles where it does abnormally well are all power hungry titles. He didn’t show the clocks conveniently lol. 😂

I am very surprised you pulled the 9070XT. I thought you be running that one for a while. Unfortunate that the waterblock didn’t do more for the hotspot. This is finished product. Card can draw 1200w max now (If needed) shunt works flawlessly. Very happy with results. Delta is 12c@600w and insanely good. I was not expecting results this good. Not to mentioning I have my flow turned way down SUB 200L/H to bleed air. This GPU is a real KILLA now. I’m absolutely impressed. Using Liquid Metal as a TIM, just like from the factory, I re-used the original gasket to keep LM spillage at bay and LM longevity of course. Using some soft tubing for now. At 600+ watts card backplate isn’t even warm. I used extra pads on the backside everywhere. The 5090FE+Block+Shunt is where it’s at for the ultimate value, in the already terrible value segment of 5090’s. EKWB has improved drastically, better quality thermal pads, and overall good quality GPU. Maybe a little over priced. But definitely Niche waterblock. But well worth it. I can surpass the RTX Pro 6000’s now with my little dinky FE. 🥳

Just got my PC up and running. 3AM now going to bed! Okay, shunt mod WORKS PERFECTLY! Voltage locks out to 1070mv and card runs 2,902Mhz default with no overclocking. Power usage is cut in half all the time. Temps are beyond ridiculous. My card was idling at 26c, and in games running 38c-40c with some quick testing.

PS: My card is shunted for 1200 watts of cable melting power. 🔥 🤣

FE cooler is back together for further use. whenever I remount it again.

Guys the 5090 FE cooler is amazing! Highest quality I have ever seen. Comes apart like it was engineered to be disassembled. Fans come out easily. Entire heatsink, everything! I carefully removed my 5090FE cooler for future use when I revert it back to stock, but it has impressed me. Even the fans are extremely high quality. Anyways, water cooled 5090FE is complete. I just have to put it in my rig, might use some soft tubing because I’m impatient. 🤣

I almost bought a Yeston 4080 Super on Newegg. It was a very good deal at the time which was the only driving factor, it was priced cheap and seemed like a good buy, and I liked the white PCB. It was not a micro GPU though.

Measurements are 8.4”x5” and about 1.40” thick, I think it’s maybe same size as the Inno 3D iChill 4090/5090.

EKWB product. 😎 I guess they are here to stay after all.

EKWB 5090FE block arrived. It looks like EKWB has finally included some quality thermal pads. They use to include that blue thermal tape previously 🤣 I may just use these!

I agree with that, testing the GPU’s on same system etc, but this is all I have. I think my conclusion is that I’m at least confident it is +42% minimum. But after seeing this and a few other 4090’s, I am thinking it is even more than that during native 4K. I played this game on my 4090 Gaming OC, and it could not handle native 4K sometimes (Mainly outdoors) (Indoors was just fine). I took the leap from 4090 to 5090 and in this title, no doubt it kills a 4090. I am beginning to seriously question review data now though. 🤔 This is a screen shot standing in similar fashion. It is about 60-64% easily. So, 42% would be extremely generous. My rig should be worst case scenario benchmark for a 5090 in games considering how freaking ancient it is 🤣. This is the RTX 4090 (99fps) This is the RTX 5090 (164fps)

@Papusan I was able to compare with a 4090 on YouTube video in Starfield on the planet Kreet at 4K native. No DLSS/ No Frame Gen. 5090 was pulling 46-55% higher fps. Same map, same graphic settings. Honestly, when I say 46% I’m being really generous. It’s closer to 55%. I’m trying to replicate the same scene, standing in same area etc lol. My GPU is at 2,632Mhz and memory at 17,000Gbps. 4090=70fps 5090=108fps How in the world Techpowerup has 22.5% is beyond me. But native shows INSANE gains on the 5090. I’ll share a video soon. But this performance jump is MASSIVE.

It is nice to see that 5090’s have all normalized in pricing! We shouldn’t see anymore price hikes since manufactures all started incorporating extra for tariffs long before tariffs even kicked in lol. The 5090 for 2K is already pushing it. Tomorrow is the day for 5090FE tear down. I ordered some different shunts for the shunt mod and I decided to just go all in on 1,200 watt lol. We think GPU’s are bad, exotic cars are so much worse and far more greedy. I recently learned that if someone wants to buy a new GT3 RS at MSRP, they must secure an allocation, in order to secure an allocation, you must purchase lots of new porches from that dealer, walking in and even wanting to pay over MSRP you’d get laughed at. They’ll usually agree to sell you a GT3 RS for MSRP, but you have to buy like 5 other random cheaper harder to sell Porsches for 50-100K+each. The advantage to these GT3 RS’s are that if someone could turn around and sell it for $400K+ right away, or drive it a few years and still make money on it. It’s amazing how there’s zero money loss on exclusive super cars! Only the select few can buy one though, and those same select few can enjoy it and sale for all of their money back and some when they want another brand-new model. This happens with many unique super cars. I wish I could afford such a life style lol. 😂 but good news is, many of these more wealthy people buying such vehicles can keep the money rolling over in to new cars without exactly losing it. There are some interesting videos on YouTube how dealers flat out say NO to anyone walking in trying to buy a new GT3/GT3 RS/GT4 RS/ Turbo S etc. it’s bananas. I do not think I could sell 4090HOF unless I was getting a 5090D HOF with 2x16 and the jumpers that allow 1KW+ 🤓

The EKWB 5090FE block has taken a long time to get. It has finally appeared in the United States in TN. So I’m hoping maybe Monday I’ll have my block. It feels like it’s been traveling all over the world lol.

Take a look at any 4080 Super review and look at the 4090. The 4090 is “Barely faster on paper” however, the internet never mentions this. You’ll see 20-35% leads with the 4090 even smaller than that of a 4090 VS 5090. I think it’s like a +27% 4K gaming average that the 4090 beats the 4080 Super by. Who in their right mind would pay +600 dollars for that tiny upgrade? EVERYONE APPARENTLY lol. So, my point is, it’s definitely more than 27%. A LOT MORE. I upgraded from a 4080 Super to a 4090. It was a massive boost to gaming performance in 4K. Well, the 5090 lead over the 4090 is even larger than the 4090 lead over the 4080 Super. So My point is, real world testing is the only way to go. 4080 Super VS 4090 =27% 4090 VS 5090 FE= 35% (Astral= +42%)

I’m seeing much larger gains than 22% in Starfield with 5090 running native. The RTX 4090 wasn’t smooth enough playing in native 4K because it was typically running 60-90’s. I played the game enough where I tried to run native 4K, after getting in the zone, I always resorted to running DLSS Quality without FG with a 4090.

I can check, but I’ll say 120+ fps without DLSS or FG in Starfield is pretty easy for 5090. Unfortunately I didn’t have any comparisons with my old 4090 using those settings at native without FG. I don’t see a huge performance loss when I turn off DLSS Quality and go to native with 5090. I think the 5090 is definitely faster over the 4090 than people make it out to be. The efficiency is really amazing as well. Many scenarios I’m finding my 5090 using the same or just slightly more power than my 4090 like maybe 25 watts more, all while being much faster. It’s not too terrible for $1,999.99. Performance is there. Efficiency is there for sure. Performance has gone up everywhere! I’m feeling the slight effects of PCIe X16 3.0 bandwidth bottlenecks with a 5090 which is about 2% at 4K and 4% at 1440P.

I compared my RTX 5090 with my RTX 4090 from a previous video that I made. The game was Starfield @4K/ DLSS Quality+ Frame Gen. 4090 Max OC@98% GPU Usage=192fps 5090 Mild OC@96% GPU usage=272fps This is the exact same save game, and the same scene on both comparisons. The 5090 is +42% faster than the RTX4090. ^ This is pretty dang good! If my GPU usage was 98% with the 5090, it would be even more. Also, the 4090 was completely clapped out on overclocking. While the 5090 is not. It’s running a pretty mild overclock. Also, almost the same power consumption between these cards!!! 🥹 I can detune my 5090fe to like 335-350 watts in this game and still hold 270fps. I am seriously impressed! I made a video on this, and will share soon. The 5090 is definitely a HOSS.

Yep, Basic/advanced/pro are all solid black thermal pads. Never seen it before with thermal pads.

These pads seem excellent. This is a new product for 2025 from Thermal Grizzly. I have no W/mk rating on these. It’s just Basic/Advanced/Pro levels. I went with Advanced. They feel very soft. I have used cheap blue Chinese 8w/mk thermal pads with amazing results before. For the price it seems very good if you buy one you really get two packs. So it’s really $12 for 100x100mm. They are black so you won’t see them against black FE PCB from the side. (always disliked that with colored pads) Also, these pads feel very soft.