2018 Hardware Thread [2]

From reddit:
Just FYI, i9-9900k went poof. Died suddenly while I was typing an email. 35c idle temps, no issue for 1-2 weeks while its been in. : intel

Woo, EVGA has a water cooled card, and get this, it`s in stock.

EVGA - Products - EVGA GeForce RTX 2080 Ti FTW3 ULTRA HYDRO COPPER GAMING, 11G-P4-2489-KR, 11GB GDDR6, RGB LED & iCX2 Technology - 9 Thermal Sensors - 11G-P4-2489-KR

Fairly nice looking I suppose

Geneo said:

I told you that about a week ago in this thread. But I said color cycle rather than breathing. Color cycle can also be done in hardware to some extent, but the cycle of the motherboard is out of sync with the memory, so it needs CPU to do it properly (you can see this by setting it to color cycle, then boot to the bios).

I know you did, and that's why I did a little further searching and experimenting.

---

Basically for now I have three choices:

Use static, which gives a wider range for the colors anyhow, by offering saturation and brightness(no need to light up the whole room), and use 0% CPU and 5.8MB RAM.

Use the Sync settings I really like-(glowing yo-yo or flash and dash) using from 4.1 up to 5.6% CPU and 6.1MB RAM.

Disconnect the Trident-Z's from Aura- using only 1.2% CPU and 6.0MB RAM, but that doesn't turn the RGB off, it runs in that rainbow-aura mode I hate soooo much(I try to make the LEDs match my current wallpaper and system color scheme).

Lastly, turn AURA totally off- when doing benchmark runs.

I wonder if it is the RGB on the Trident-Z's that make it near impossible to make any changes to any RAM settings(other than raising the speed through BCLK) and be able to boot???

OldMike65 said:

You are so dam lucky to have that chip Cliff.....as low as I can go in Offset mode for 5.1 just to be stable is 0.395v

Cliff S said:

I just raised it only 0.005 instead of 0.010 like I said above, and it was enough to push it over to a steady 1.168 with no voltage fluctuation(It's better for the VRMs that way).

Not so dam lucky Mike
Under extreme benchmark loads it ran good.

But this morning(German time) I had a couple of BSODs during idling(just the browser open and what ever background tasks), and they seem to have happened so fast that NO crash dumps was created both times

Since the crash dumps are made from what is in memory, I tested my memory first using MemTest64 at maximum for 10 loops, which runs inside of Windows, and finds errors faster than the offline memory test software. but memory passed the test.

I then decided to run IntelBurnTest at Standard and the CPU passed.
So I tried Burn test at Maximum and after 4 minutes Windows BSOD'd, so I upped the voltage a little, but again BSOD.

Finally I decided to set my voltage back up near where it was originally, at 1.376V, and let the burn test run at maximum again, my CPU passed, temps got a bit warmer, into the mid 80's sometimes, instead of mid 70's, then with the lower voltage, but after 3 hours & 5 minute it completed.

So I guess I'll use 5.0GHz @ 1.376 as my daily(AVX load set to 0).


Christ all that work yesterday for the cats(but it was fun, and could still bench mark)

Cliff S said:

Not so dam lucky Mike
Under extreme benchmark loads it ran good.

But this morning(German time) I had a couple of BSODs during idling(just the browser open and what ever background tasks), and they seem to have happened so fast that NO crash dumps was created both times

Since the crash dumps are made from what is in memory, I tested my memory first using MemTest64 at maximum for 10 loops, which runs inside of Windows, and finds errors faster than the offline memory test software. but memory passed the test.

I then decided to run IntelBurnTest at Standard and the CPU passed.
So I tried Burn test at Maximum and after 4 minutes Windows BSOD'd, so I upped the voltage a little, but again BSOD.

Finally I decided to set my voltage back up near where it was originally, at 1.376V, and let the burn test run at maximum again, my CPU passed, temps got a bit warmer, into the mid 80's sometimes, instead of mid 70's, then with the lower voltage, but after 3 hours & 5 minute it completed.

So I guess I'll use 5.0GHz @ 1.376 as my daily(AVX load set to 0).


Christ all that work yesterday for the cats(but it was fun, and could still bench mark)

That is kinda funny because when I am @5.0 I use the exact same volts, in Manual mode. (1.376v) AVX always at 0
Offset does not seem to be 100% stable for me, when benching.

OldMike65 said:

That is kinda funny because when I am @5.0 I use the exact same volts, in Manual mode. (1.376v)
Offset does not seem to be 100% stable for me, when benching.

Oh that reminds me, I also switched back to manual from offset.
Basically my old settings.

Cliff S said:

Oh that reminds me, I also switched back to manual from offset.
Basically my old settings.

Your chip still oc's better than mine....I can run @5.1 stable. (1.410v)

Edit: I'm using 1.400v @5.1 Manual mode, with AVX at 0

I could put my multiplier up to 52(5.2GHz), and set AVX load to 2, and voltage up to 1.392 and be stable(that was my setup for a few weeks) but then the AIO's radiator fans need to work a little harder.
At 5.0 Thermaltakes UI says the water temp is only 34 to 37°.


At 5.2 GHz water temp hovers in the mid 40°s

Been benching for last couple hours with no bsod's what so ever too.