jones177

To get consistent scores in Time Spy it is best to run your GPU fans at 100% until it gets to around 30c. This will take the heat out of the case and give more consistent scores. 11200 is what you get at 4.9ghz or it is basically what you get using MCE on a Gigabyte board.

I did test Battlefield 5 with my Strix 3080 paired with my i9 9900k. The frame rate was the same at 1080p and 1440p so CPU bound at those resolutions. At both resolutions The GPU was only at 60 and 65%. At 4k the GPU was at 96 to 98%. The frame rate at 4k ultra was around 80 to 95 and at 1080p it was on the 200fps limiter. I did test this setup against my i7 8086k with a 3080 ti and it was faster at 1080p and 1440p but lost at 4k. If you don't have a 5000 series Ryzen or a modern Intel CPU you are not going to get the performance out of a 3080 at 1080p or 1440p in games like Battlefield 5.

My issue is the heat and with that the sound on an air cooled GPU. At 250 watts the GTX 1080 ti ran cool in cases that were built with GTX 680s in mind. With the RTX 2080 ti at 300 watts stock I only had to add 1 or 2 fans to the case to get the same temperatures as the 1080 ti. With the 30 series I have cards that use at stock between 350 watts to 400 watts and had to re case my gaming PCs to get close to the same temperatures as the older cards. I think we are at the limit now when it comes to air cooling so undervolting, repadding and repasting has a thing now. We are basically reengineering the cards to make them work as they should out of the box. Yesterday Newegg had a FTW3 Ultra 3090 for sale at MSRP and I wanted it. The price did not stop me but the stock 420 watts did. The 400 watts is my limit with the 30 series and unless they come up with a new way to cool GPUs it will be the limit for me with the 40 series as well.

My place is a constant 21c and I get 90c on the GPUs at most. Before retiring my computers were kept at 16c. I had to wear a pullover to work. It was sort of crazy in the middle of a Florida summer.

Heat has been annoying with all my 30 series cards except my Stix 3080 with its silly big cooler. For 3090s I went with MSI Gaming X Trios that have heat pipes on the back plate. The vram stays as cool as the 3080 ti version. They are not much fun since they are severely power limited but for rendering at stock they do fine.

Why annoying? I use 3090s and I have not come across anything annoying.

That is about right for BF5(ultra) at 1080p with a 3080. I played with my i9 9900k and Strix 3080 it was about 35 to 45% CPU and 60% GPU. I now have a 3080 ti with the i9 and the GPU percentages are the same but the frames are slightly higher(nailed at 200). Even at 1440p I am still at 60% on the GPU since the game is CPU bound. Only at 4k did I get 96 to 98% on the GPU.

I won't throw them out and if GN does do anything on AIOs again I will contact them.

It depends on the game and the resolution. My i9 9900k is paired with a 3080 ti and at 1440p it does well. Here are 2 games that I have tested and in one even at 1080p the i9 9900k does about the same as newer CPUs. The i9 9900k is using MCE. The i9 10900kf is using TVB. The Ryzens are stock. Here is Horizon Zero Dawn at ultra settings. 1080p 1440p 4k i9 9900k(8 core) 155fps 138fps 88fps i9 10900kf(10 core) 157fps 138fps 87fps 5800x(8 core) 156fps 138fps 89fps 5900x(12 core) 156fps 138fps 88fps Shadow of the Tomb Raider with the highest preset. 1080p 1440p 4k i9 9900k(8 core) 169fps 154fps 96fps i9 10900kf(10 core) 189fps 164fps 97fps 5800x(8 core) 198fps 163fps 96fps 5900x(12 core) 211fps 175fps 105fps This is the i9 9900k vs the 5900x in Assassins Creed Odyssey on ultra. 1080p 1440p 4k i9 9900k 106fps 89fps 72fps 5900x 104fps 93fps 69fps In games that use more cores or like IPC over frequency the i9 9900k can lose big time but in most games it doesn't. In actually game play the more modern CPUs do better in Total War Warhammer II at 4k ultra but that is about it in my experience.

They only have a one year warranty. When I bought the first one there was a shortage of AIOs so if I wanted a top brand I would of had to pay scalper prices. I took a chance and it did not work out. I got the second one because the first on did well cooling a i9 10900k. It was not as good at cooling as the first one so I knew there was a quality control issues. When It got loud I just wanted it gone. The AIO head has 6 torques and 2 custom screws. When I find a head that fits the custom screws I will take it one apart

I have a White build with a Corsair 5000d and a Strix 3080 white OC but it uses a black aio. If I went white on the AIO I would tray a Lian Li Galahad White.

I would not buy a Silverstone AIO. I spent part of yesterday replacing one. It was bought in July 2020 and the pump failed. It was replaced by an EK. Another that I bought in August 2021 became unusably and was replaced as well. It started to underperform and became loud. I also had 2 Silverstone case fans fail as well that were in my i7 10900kf build. So in my last 5 builds the only failures I have had is with Silverstone products.

I use 60mm Noctua fans below the GPUs on my cases that have PSUs on the bottom. I started using them with RTX 3080 ti since they put a lot of heat into the cases. On my CM H500P Mesh and CM H500 ARGB they are good for 3c on the GPU but only 2c on my Corsair 5000d. My main gaming PCs now use Lian Li o11 Dynamics that have all the intakes below the GPU. They fix the issue of hot air getting trapped below the GPU but make the air that the CPU cooler intakes hotter. Not a big issue with a 360mm AIO but could be with less.

With my 2080 tis that used 300 watts stock I had to add fans to my cases the were fine with GTX 1080 tis(250 watt stock). My RTX 3080 ti/3090s use between 350 and 400 watts stock and I had to re case my 2 gaming rigs just to get the same temps as the 2080 tis. I have 2 cards that can do 450 watts and in tests without the fans at 100% it is not doable in a case. The way GPUS are designed has to change if 450 watts stock is doable other than an open bench.

Yes. All the GPUS I have listed below have the vram at those temps under load.

Here is a good video about the subject.

My Strix 3080 is silly cool even using 450 watts and even the vram was cool at +1200. It is defiantly a fun card. At stock my cheap EVGA XC3 3080 ti beats it and that is why it is not in one of my gaming PCs.

I have a Srix 3080 and it is built a lot better than the EVGAs this series. I had the Stix in several cases and it put out about the same amount of heat as my old 2080 tis even using 70 watts more. If I buy another 30 series it will be a Strix. My EVGA FTW3 3080 ti is the card most likely to be RMAed. It made all kinds of strange sounds in the menus of Total War Warhammer III. The only thing EVGA has going for it now is they are cheaper and easy to RMA.

My builds using them are too new to have dust issues. The oldest one is about October 2021. Time will tell.

On one I used the little zip ties you get with motherboards. On another I used 2 fan screws since the holes in the PCIe slot covers were big enough for the screws to fit.

The CM H500M(your case) works but the H500P Mesh doesn't. I bought the H500P Mech after the H500M since the M worked so well. It only cost $150 compared to the M at $235. Since it has the same layout as the M I thought it would perform about the same. It didn't. Here is what it looks like. It was fine with a FTW 2080 ti but with a XC3 Ultra 3080 ti it gamed in the 80s. Hot air form the GPU was getting trapped below the GPU and was not escaping the case fast enough. This also happens with the H500M but to a lesser extent. Just touch the glass below the GPU after a gaming session. A 60mm Noctua fan installed below the 3080 ti got it to game at about 75c(image below). When I tested the case with the MSI 3080 ti it was back in the 80s gaming. I think it was the difference in size of the GPU that made the difference(XC3 Ultra 285mm X 111mm vs Gaming X Trio 324mm X 140mm). I replaced the H500P Mesh with a Lian Li o11 dynamic.

Your old score is at 62c and your new score is at 64c. If your GPU downclocks at 64c that could make the difference. It is not a 4k bench so CPU performance matters more that what it would using Time Spy Extreme. The bench also likes CPU frequency so if your CPU is off a bit it will reflect in the score.

I bought a second Gaming X Trio 3090 last week. It is basically a spare for the 3090 in the H500M(used for video editing). A Suprim X would be a lot more fun but it has to work in the H500M as is since that computer can't go down for even a day. I do consider the Trio conservative and power starved with any overclock so not fun at all but it is cool and quit in the H500M. I may get a 3rd 3090 and that can be a fun card so it will be either a Suprim X or a Strix. My AIO top fans on the H500M are intakes since the CPU runs cooler that way. I have an i9 10900k using Thermal Velocity Boost in it so it has to stay below 70c. With it exhausting it was pulling up hot air from the GPU and the CPU was in the 80s. I set this up when the computer used a 2080 ti but it still works with the 3090. My stock H500M is much better at cooling a 3090 than my Corsair 5000d with 11 fans. You could replace the exhaust fan with a high performance Noctua but it may not do anything at all. It did not with my Corsair 5000d. My CM H500 Mesh was a fail with even the MSI Gaming X Trio 3080 ti(350 watts) so I did look into replacing the 200mm fans but after reading this I decided to get a different case. https://www.gamersnexus.net/hwreviews/3161-noctua-nf-a20-200mm-fan-review-vs-cooler-masterfan-mf200r GPUs can sustain higher clocks at cooler temperatures so as they get hotter they drop clocks(mhz). It is why water cooled GPUs run at higher clocks than air cooled ones. Do a bench and set your fans at 100% and watch the mhz go up as the temps go down.

I think your CPU was having an off day. Here is my 24/7 overclock test. https://www.3dmark.com/3dm/64110353? Even though my GPU clocks are higher than your 2021 test it is let down by my i9 that was underperforming.

I had a Strix 3080 on my i9 9900k rig for a while. This is what it did stock. https://www.3dmark.com/3dm/63945593 On your high score you are averaging 2c cooler. The card downclocks at around 64c so that can make a difference.