Overclock.net banner
3,041 - 3,060 of 18,210 Posts
yeah , and stores can’t do a single thing regarding the bots. Talked to Proshop today, they have a delivery incoming , 1-5090 TUF OC and 3 Astral OC for Scandinavia and Germany…
Proshop tried to tell me that they have systems to prevent bots...

🤣

So here we are again. ngreedia is going full r-e-t-a-r-d holding back cards to hype them up and raising prices, resulting in people who don't need cards buying them with bots to make money.
 
  • Haha
Reactions: Papusan
Proshop tried to tell me that they have systems to prevent bots...

🤣

So here we are again. ngreedia is going full r-e-t-a-r-d holding back cards to hype them up and raising prices, resulting in people who don't need cards buys them with bots to make money.
Yea Nvidia has quickly moved to the top of my most hated tech company list. Lol.
 
Microcenter is my only hope and it's 170 miles away. Even my turbo buggy BMW can only get there so fast. LOL I do have a good friend that's 15 min away. Last resort I'll pay him to snag one.



Do you still have the ECC setting in NCP?


Some of that is run to run variance, but look like it helps up to around +1000 and then regresses. I don't think that's ECC, more likely just not helping and robbing a tiny bit of power from the core.

Oh and try Speedway if you want to look for better memory scaling. It's heavily memory bandwidth bound on the 4090.
Are you in california?
 
god i cant stand him (jayztoscents)
 
  • Rep+
Reactions: hazoR94
Technology Advertising Graphic design Electronics

Well I have found out where one 5090 Astral is in Australia. It is in this Custom Build "ready to go" AUD price converts to $7100 USD
 
Proshop tried to tell me that they have systems to prevent bots...

🤣

So here we are again. ngreedia is going full r-e-t-a-r-d holding back cards to hype them up and raising prices, resulting in people who don't need cards buying them with bots to make money.
Yeah , they told me the same. That’s total BS 😂
 
As i saw from the first tests, the 5090 downclock to 2900+- average under Timespy. Why don't lock it at 1v max voltage and go for the maximum stable clocks?

Furmark is not dumb at all, it's a part of testing if the gpu's vrms are stable and if there's too much coil whine.
You can chekout the coilwhine with highfps/highload (with a lower resolution) or lowfps/highload (with maximum res).

OCCT works the same way at higher shader complexity.
Saying that stresstests are dumb isn't exactly a smart thing to say, to put it mildly...
so testing vrms is useful for? testing coil whine? ok

is dumb to complain that in a furmark scenario the clocks are dropping from an average game workload....as simple as that

what does HOF mean? thank you
that's galaxy oc enthusiast vga series
 
  • Rep+
Reactions: spin5000
Some people on reddit were complaining about performance degradation after system reboots. Their 5090s were clocking over 3000mhz boost on the regular, then they did a system restart, and now they are only getting ~2700mhz boost on average in the same games/benchmarks. The poster I talked to confirmed not touching voltage on MSI afterburner. Another anecdotal post stated that MSI afterburner is not always applying OC settings correctly in general.

Anyone experience these issues so far or something similar? My 5090 FE doesn't like to boost above 2975mhz or so under heavy load and will occasionally crash in games/benchmarks if it gets around that area (+250 core OC with 104% power). 2950mhz and below is rock solid stable. The card does however occasionally boost to 3100-3150mhz under very light loads for a split second which is interesting and kind of concerning though. I haven't played with voltage yet.

With 3dmark benchmarks, my clockspeed tends to stay around 2700mhz-2800mhz area and never hits the 2900mhz area.

Any idea why we are seeing such wide swings in clock speeds? Is it the low power limit?
 
Some people on reddit were complaining about performance degradation after system reboots. Their 5090s were clocking over 3000mhz boost on the regular, then they did a system restart, and now they are only getting ~2700mhz boost on average in the same games/benchmarks. The poster I talked to confirmed not touching voltage on MSI afterburner. Another anecdotal post stated that MSI afterburner is not always applying OC settings correctly in general.

Anyone experience these issues so far or something similar? My 5090 FE doesn't like to boost above 2975mhz or so under heavy load and will occasionally crash in games/benchmarks if it gets around that area (+250 core OC with 104% power). 2950mhz and below is rock solid stable. The card does however occasionally boost to 3100-3150mhz under very light loads for a split second which is interesting and kind of concerning though. I haven't played with voltage yet.

With 3dmark benchmarks, my clockspeed tends to stay around 2700mhz-2800mhz area and never hits the 2900mhz area.

Any idea why we are seeing such wide swings in clock speeds? Is it the low power limit?
Yes power limit is the problem(y) your voltage cant stay at max with hight frequency so it's around 1.0v and frequency back to 2800.
 
Do you guys think that 5090 availability will improve any within the next month? I'm filing our taxes early this week and expect to get some money back within about 3 weeks maybe. I was going to use my portion of the money to purchase a 5090 and two more EK Quantum Surface X series radiators (thickest models) (58mm thick).

I have room for two more EK Quantum Surface X360M radiators up at the top of my Corsair 1000D (side by side). These new EK radiators are super low in coolant restriction and should only need the one D5 pump in my loop to circulate coolant through a total of 4 internal radiators. I can always add a second D5 pump if I need to.

I currently have (2X) EK Quantum Surface X480M radiators sitting side by side at the front of my case, Optimus Signature V3 AM5 water block, EK ZMT 10/16mm soft black EPDM tubing.
My recommendation? Forget internal radiators and get a Mo-Ra 4. Much easier to work with. Quieter. Better for future builds and gives you freedom of choice in case selection. Coming from someone with 2x 480mm and 2x 360mm also in a Corsair 1000D.
 
Does anyone with dual monitors or more use a 5090?

I've noticed that if I connect only 1 of the 3 monitors I use via DP I don't get black screens and everything works perfectly, once I connect 2 or 3 monitors all for DP the problems start.

What do you think about the graphics score?
I use 2 HDMI and 1 DP with no issues.
 
  • Rep+
Reactions: RaMsiTo
The RTX 5090 is definitely powerful enough to handle Cyberpunk at 4K RT Overdrive preset with DLSS Quality, achieving 60-70+ FPS in my testing without the need for Frame Generation (FG).

However, I've noticed a bug or significant performance drop when enabling shadow play, making the 60-70 FPS feel sluggish during gameplay.

Switching to Balanced mode resolves this issue. So, I decided to use x4 FG instead. It feels fine to me, but if I wasn't recording, I'd switch to 2x FG or simply run DLSS Quality with PT.
RTX 5090 Gaming Trio Benchmark | Cyberpunk 2077 Path Tracing | 2160P Maximum Settings DLSS Q + FG x2

I'll also add that the video play back does not appear completely smooth, although it felt smooth whilst playing. The frame time graph is also completely flat indicates smooth performance.
 
The RTX 5090 is definitely powerful enough to handle Cyberpunk at 4K RT Overdrive preset with DLSS Quality, achieving 60-70+ FPS in my testing without the need for Frame Generation (FG).

However, I've noticed a bug or significant performance drop when enabling shadow play, making the 60-70 FPS feel sluggish during gameplay.

Switching to Balanced mode resolves this issue. So, I decided to use x4 FG instead. It feels fine to me, but if I wasn't recording, I'd switch to 2x FG or simply run DLSS Quality with PT.
RTX 5090 Gaming Trio Benchmark | Cyberpunk 2077 Path Tracing | 2160P Maximum Settings DLSS Q + FG x2
At the 17s mark in the bottom-left side where the papers or whatever are flying around, is that full disappearing/reappearing flashing effect from frame gen?
 
At the 17s mark in the bottom-left side where the papers or whatever are flying around, is that full disappearing/reappearing flashing effect from frame gen?
Could be, I didn't personally notice any corruption whilst playing but I wasn't particularly looking for it. When I was testing without recording, you could see some FG artifacts in fast movement. I found them much less noticeable with x2, which is my preferred choice when using FG.
 
The RTX 5090 is definitely powerful enough to handle Cyberpunk at 4K RT Overdrive preset with DLSS Quality, achieving 60-70+ FPS in my testing without the need for Frame Generation (FG).

However, I've noticed a bug or significant performance drop when enabling shadow play, making the 60-70 FPS feel sluggish during gameplay.

Switching to Balanced mode resolves this issue. So, I decided to use x4 FG instead. It feels fine to me, but if I wasn't recording, I'd switch to 2x FG or simply run DLSS Quality with PT.
RTX 5090 Gaming Trio Benchmark | Cyberpunk 2077 Path Tracing | 2160P Maximum Settings DLSS Q + FG x2

I'll also add that the video play back does not appear completely smooth, although it felt smooth whilst playing. The frame time graph is also completely flat indicates smooth performance.
I'm curious about how do find quality with 4x MFG? Are there noticeable issues during gameplay?

Edit: nevermind you just answered it practically.
 
3,041 - 3,060 of 18,210 Posts