[Sammelthread] Offizieller AMD [[ RX6700 // RX 6700XT // 6750XT // X6800 // 6800XT // 6900XT // 6950XT]] Overclocking und Modding Thread [[ Wakü - Lukü - LN2]]
- Ersteller xXx-Eagleone-xXx
- Erstellt am
@Johnson73 mehr punkte bei weniger takt?
Sylwester.
Experte
🏆🥇
drstoecker
Experte
- Mitglied seit
- 22.11.2015
- Beiträge
- 871
DieDreiBeiden
Experte
Der hat sich wohl hier im Forum ein paar Sachen abgeschaut
Sylwester.
Experte
well yesterday I decided to use the natural conditions to play with the computer , but I was not able to beat the previous score , there were a few runs in the range of 23700-23800 points but nothing extraordinary
1.
I tried to increase the soc clock to highest values - FAIL (at values of 1800mhz even windows did not get up , and what's strange behaved as if the disk died , I have 2 systems "windows" with the selection at startup, and did not even get to the selection screen, only messages like , automatic repair, system restore, etc. etc. I was already nervous that I broke the disk - condensation or something similar, but entered into emergency mode and deleted the settings mpt,ssd returned to work)
2.
after that it's time to play with flck, but here momentarily black image, 2250 is the maximum I can set
3.
ok, then further reducing the voltage of the memory(I've been reducing the voltage of the memory because it helps to achieve more stable runs) but here I've also reached the minimum , if i set lower values for memory voltage , runs are not stable , fps jumping
4.
then an attempt to change the minimum values of VOLTAGE CONTROL PARAMETERS in the tab POWER AND VOLTAGE - no result with raise min value , reset the driverr after reducing these values
5.
then time for Vmin Low/High (mV) values ( all 4 values), Temperature Dependent Vmin window, POWER & VOLtage tab
5a.
trying to lower it, driver reset
5b
but when I increased it (1000mV) then
Boooooooooooooooom
it passes the test on gpu clock 2649mhz where before it was failing "test 2" every time, now only sometime failing, ok I increase the clock to 2653 again fail every time, so voltage on 1050mV first gear fail, quickly start next gear passed
, so I clocked up again but braver 2660, nothing - fail each time, so the voltage on the rich 1150mV, EVERYTHING OR NOTHING ;D
test passed !!!
than increase the clocks to 2670 and - GET the RECORD, I was as excited as a dog at the sight of a ham , I started taking pictures, posting on the forum
, this way I fought for 5 hours , i was already falling on my face, so I did another test on 2675 clock, but there was a failure, I increased the voltage to 1175mV - because I was more afraid - this is more than the allowable 1150mV for GFX voltage, but also did not help, and I gave up until the next attempt, I believe that this is not the end , WINTER IS COMING
Mal abwarten.Wir wissen alle was TS da an Taktraten ausliest.Hoher Coretakt und und so ein niedriger Durchschnittstakt,
Denke aber auch das 26.500 möglich sind.
Der hat aber auch eine gute CPU.
Weil zur Time Spy Rangliste viel zum Einfluss der CPU-Leistung zu lesen ist, hat sich euer Fleißtiger Arbeit aufgehalst und stellt eine 3DMark Port Royal-Rangliste zur Verfügung.
CPU-und RAM-Leistung sind in diesem Raytracing Benchmark nicht ausschlaggebend. Viel Spaß!
CPU-und RAM-Leistung sind in diesem Raytracing Benchmark nicht ausschlaggebend. Viel Spaß!
Zuletzt bearbeitet:
Thanks @Sylwester. These are valuable insights, thanks for sharing.
ILoveTattoo
Enthusiast
Stell mal bei HW info die Aktuallierungszeit auf 50ms und lass mal TS laufen. Wette das dein Peak bei ca 700-800 Watt liegt im GT2. Eventuell sind die Lastspitzen einfach zu hoch
Sylwester.
Experte
no problem, if anyone ever wants to do a joint record-breaking session, we could get together, connect the computers via teamviewer and get voice communication via whatsapp, maybe together we can work something out
Danke guck ich mir später mal an aber auch diese Spitzen müsste doch mein neues 1200Watt NT halten können, oder nicht?
Also mal im ernst, ich hab die lc und die toxic mit einem 850w gejagt und da hat sich nichts abgeschaltet.
Neues Netzteil und gleicher Fehler, da kann ich einfach nicht mehr an die psu glauben.....eher an eine zu hohe last auf dem pcie slot. Weil auch dort wird die last über die specs gehen wenn du aufdrehst.
Neues Netzteil und gleicher Fehler, da kann ich einfach nicht mehr an die psu glauben.....eher an eine zu hohe last auf dem pcie slot. Weil auch dort wird die last über die specs gehen wenn du aufdrehst.
ILoveTattoo
Enthusiast
Wen man die Daten mitloggt , zumindestens mit dem Evc liegen die Spannungsspitzen deutlich drüber , das dann die Schutzschaltung des Netzteils auslöst . Igor hat Mal vor langer Zeit ein solches Video dazu gemacht .
Sylwester.
Experte
Are you sure it's the power supply? You have already replaced the power supplies and the problem persists, check the temperature of the chipset on the motherboard, check whether the card is not obstructing the fan on the chipset, if there is one on your motherboard, and for sure , just for test move the card to another slot on the motherboard.
Sylwester.
Experte
throw away that riser, I bet it's the one causing problems, or at least try to replace it with another one, it's not a coincidence that the full pcie x 16 connector in 4.0 or even earlier 3.0 standard is located just below the processor the shorter paths the better signal quality, the signal flowing there and the amount of data is very sensitive to interference, and you add another few lengths, so you probably also have worse results, because more data is subject to error correction , after all, for example if the sound card on the board is poorly shielded , then you can hear in the speakers a beautiful distortion in the sound , ,The interior of the computer is like a small microwave, the interference from the magnetic field of coils in the power supply system of the motherboard or graphics card spread throughout the build, and you add to this field your riser, which passes a huge amount of data during the game
Zuletzt bearbeitet:
Sylwester.
Experte
some times You can get problems , smoe times not, depeence of quality of the riser , how is mounted in teh case , and many many other things , are You shure that not affect to Your score? did You try make test without this ?
There's a PROBLEM with Riser Cables
Riser Cables and vertical GPU brackets are getting more and more popular for gaming PCs Let's take a look at potential problems an how to fix them
hardwarecanucks.com
someetimes is no matter about speed but sometimes :
and for example in 3dmark ts difference adotu 1 or 2 fps is like 300 or 500 points less on gpu score
gpu score difference 130 points , test one difference 0,4fps test 2 difference 0.9 fps
Result
www.3dmark.com
Zuletzt bearbeitet:
Was willst du damit, wenn dir die Kiste bei 500W ausgeht? Ich würde erstmal versuchen den Fehler zu finden.
das ist nur bei einer bestimmten Anwendung passiert, nicht 3Dmark, da mach ich mir keine Sorgen.
Beitrag automatisch zusammengeführt:
Schon gefunden, welchen sollte ich davon nehmen, @L!ME könntest du bitte mal schauen, danke.
Amazon.de : CH341a
www.amazon.de
Zuletzt bearbeitet:
