Du verwendest einen veralteten Browser. Es ist möglich, dass diese oder andere Websites nicht korrekt angezeigt werden. Du solltest ein Upgrade durchführen oder einen alternativen Browser verwenden.
you know that's the thing about manners, they are not written in a rule book however still being adhered to by the regular folk on here. Your reply makes it seem like you either do not understand that concept or you just choose to be a cunt. This is not OCN.
Since I started unfortunatelly its my job to try fix it. I apologyze if i was bit toxic.
Discussing is the hardest thing in earth.. People will have confidence in their thoughts, defend their claims, have personal attachment etc.
Some people "refuge" is this forum, even not being german. Ppl are here to help each other without toxicity and still with knowledge probably cause other places had people without patience/respect.
I want you to not go away, i want you to see that i apologyze if i was not gentle with you.. But the intentions are to help!
DDR5 is different from ddr4, is much harder, is way more complex and its NEW
Even the smartest people are often wrong!
If i say to you to start again from scratch and with lower expectations, is cause the road is LONG.. ddr5 learning is step by step
requires ton of time. i took 4 months to kinda step up and im still clueless about so many things..
again, there is NO SHAME in using lower frequencys.. its actually.. respectable if u do it fully stable..
there is not a single scenario in EARTH
that
tRDRD_dg 8
tWRWR_dg 8
tFAW 32
tRDRD_sg 8
will hurt any test, dont make sense to stretch the unnecessary for testing
otherwise people would be recommending test frequency at "Load RAW mhz"(which is like 126+ all timings+)
QXE said once that ISNT good to test super relaxed stuff aswell.. you just type something safe...
u can safely put:
tRDRD_dg 8, tWRWR_dg 8, tFAW 32, tRDRD_sg 8, tRDRD_sg Auto, tRAS 60, tRTP 12, tRDWR both 20, at 7800 ez
That's not what I asked you to explain. I asked you to explain the implications of having the timing that you singled out as bad, tRDRD_sg = 16, for a scenario where I am experimenting with how far my system will push ram frequency. If you cant explain that then its completely pointless you linking that Buildozid video as he explains why the setting is bad in a benchmarking scenario, which isnt my scenario. Maybe you are not so good in English so are confused, I don't know.
The timings I have entered are safe and are super relaxed so what's the problem?
Doesn't make sense to prescribe settings to someone as safe when every system reacts different to different settings. Might work for your system but not mine. The whole point in tuning settings is to come from settings way out and drop each 1 by 1, getting tighter and tigher, and testing stability each time. Its called being methodical. If you add in a bunch of tight settings at once and its unstable, how will you discern which setting caused the instability?
tRDRD_sg higher than 8 is just a sin
thats why i said no sense, cause 8 will *always*(im 99,9% certain) work and its important for performance(as per bzoid vid)
I also spent a few minutes trying to help him with these hints, 12 hours of streaming and expanding his frustration with Intel. It kept raising voltages, and his starting point was always wrong. I didn't know Madness was also there.
An dieser Stelle möchten wir Ihnen ein Youtube-Video zeigen. Ihre Daten zu schützen, liegt uns aber am Herzen: Youtube setzt durch das Einbinden und Abspielen Cookies auf ihrem Rechner, mit welchen sie eventuell getracked werden können. Wenn Sie dies zulassen möchten, klicken Sie einfach auf den Play-Button. Das Video wird anschließend geladen und danach abgespielt.
Youtube Videos ab jetzt direkt anzeigen
Never give up~~
Whatever was, is past
See you made it further than before.
You can't see how it will be tomorrow,
but you can say that its better than yesterday and seek on working hard
Your choice is only to continue with whatever happens.
Things change in life, same as in tech. They break, to make room for a better new start
Same as plants vanish to soil, soo something better can grow with past experience and become more beautiful.
Its not the event that sets the tone. Its only on you what you do with it~~
Don't become the soil, stay a strong growing flower 🌸
Keep on working hard with the head up to the sun! 💪
Thank you for the kind words and motivation quotes.
No pain, no gain!
I flashed the 9901 which wasn't able to do 8533 at all before, but I tried it with low SA (1.12V) and it will work, it will need a small adjustment here and there.
I went back to LLC4, so maybe the V/F needs little adjustments too and not the RAM at all. I will test that later as well.
The profile is progressing nicely. I would like to note one thing, to ask you to take the photo under load. It shows more useful information than the idle state.
The profile is progressing nicely. I would like to note one thing, to ask you to take the photo under load. It shows more useful information than the idle state.
I'm not only interested in Vcore, I also look at the Uncore VID values, for example. And many more data.
I tried several Curves and at the same frequency, same Vcore, the Uncore VID changes.
The V/F Curve is very important, especially in a case like what I'm doing, that I don't let it go until the end.
Yesterday, not only was a new Curve made, but I also reinstalled the cpu. It is certain that the PC behaves completely differently now, in a positive direction.
As you know, the socket was dirty, I also cleaned it very carefully. That's how I bought the motherboard used, now that I rebuilt the water cooling, I also did this.
All is fine 🤭
But those "refuge people" come because like you mentioned, "may find information here that is not on their forum"
Some information sharing people have spoken german for 23years, but may be slightly si"k of it or just got used to the language that the tech is developed at. That is english, given chinese is hard.
Some things are easier to explain that way, and it seems respected. I hope;
Forum still is a german one, with an exception of english.
Mein bestes stabiles war 7400MT/s auf einem 12900K, mit viel arbeit ohne DFE.
7600MT/s ging dann nicht mehr auf dem gutem testers-sample.
8200MT/s kaam noch ins Bios, jedoch war bei 7400MT/s realistisch schluss. (gutes Z690 APEX auf Z790 Bios+personal tweaks)
Der Tester bekaam 8000MT/s durch auf derren Z790 Dark?, bzw dem Z690 Unify-X.
Für meine normale Kühlung, ging über 7400 kaum bis garnicht.
Dafür ist 12th gen eine zu große Diva. SignalIntegrity war einfach zu schlecht. Das Board war super !
What is your current progress @ last stable that passes y-cruncher ?
Did you start with 7600 without stability tests ?
Please keep us up to date, maaybe something can be done.
But i wouldn't realistically expect much more than 7200 at best.
On the other hand, i was a big novice 🐣and now things look very differently when understanding (mostly) VREF.
Lets see~
Hmmmm
Because voltage doesnt scale higher.
SA changes as "side-issue", the ODT around.
You can notice its a cpu issue and not a mem issue
Then you can notice IA supply does no change, and its not ring.
At this point its just signal strength, aka ODT ~ IF you trust in the board doing good RTT & ODT. Which it should now;
De-ionized (girlfriends?) Hairdryer Air
And a de-ionized (in pharmacy , medical injection fluid) purified water bath.
Then at the end 99.9% IPA and we are done.
// here its rather used to push away any potential waterdrops, even if fluid is supposed to be non conductive.
I like to use such type of clean water for watercooling mixtures, where concentrate is being used.
So that includes Alu+Copper AIOs or Brass+Nickel.
// because normal distilled water, nor purified is clean enough and if working with chemicals, i prefer to not have trust on parts i can actively mitigate
// Hence labor/medical grade water that is fitting for syringe injection or medicine distillation. Also easier to get something of quality, vs trust in a random supplier.
Alternatively C&C Red can is often used to remove electrical corrosion
tho i started to like the WD40 green! , professional lineup.
Green electrical cleaner can + a strong hairdryer
Both of those including IPA, may dissolve it and the leftover flux, but it will be left back as a residue, if you dont blast it away.
On the other hand, if you use a too strong can of similar fluids, you will damage the spring-pins and the WD40 red-can, is an oily base. Also bad !
I flashed the 9901 which wasn't able to do 8533 at all before, but I tried it with low SA (1.12V) and it will work, it will need a small adjustment here and there.
I went back to LLC4, so maybe the V/F needs little adjustments too and not the RAM at all. I will test that later as well.
It is being said that the G4 XOC bioses have a changed V/F curve and there is a warning what to use and which switch not to use.
I dont understand the warning, and dont trust the bios.
Given it uses 1002 as a base, and even 0081 is a good Gear4 overclocker ~ maybe don't risk your CPU life.
It will not gift you anything except potential Gear2 trouble. Our higher clock is hard work, not a bios issue.
Im happy that he spread good faith.
But it must have been hard, even for him.
Madness is not the most, lets say "patient" person on earth when it comes to arguments 🤭
Oh well, its also nice to see that many people try to help BZ.
I hope he shows a bit more weakness, and remains humane.
Its hard work, it never was easy. Lack of papers doesn't make it easier at all.
Me too, i have faith, but he insist on single-timing changes.
Thats maybe his and my 2-3 years ago flaw. When i trew sh*t at the wall and inspected what sticks. All sorts of shenanigans.
No timing goes alone. They all go in pairs or in trio's.
Starting to switch the perspective that way,
will let you learn from your data easier, given every change you do has an influence on something that comes before it & after it.
If it wouldn't be, then there is no reason why more talented engineers haven't implemented this as "basic specifications" for other Partners to work on
There are always some downsides to your changes.
I have faith.
He's a cool guy and its interesting to listen to the long streams.
Tho i've lost interest, the times it was an interesting topic ~ i did enjoy his long streams/videos.
It is more tame than FC's Video's which are full of action and overlapping data.
BZs are more tame and easier to process.
VDDCR_IA supply feeds the most important parts of the CPU
Cache, Ring, Cores, E-Cores, IMC , FCLK
All take a chunk of the supply ~ loadbalanced by IVR or VID state/method.
VDDCR_SA is the substrate/GMI? supply.
IA and SA in the old days had to follow a (delta, margins) pattern of 50-75mV.
This is now done automatically.
rOdt and vOdt are automatically generated too, same as the V/F curve is generated from the fused by the fab curve.
SA "set point" , which is VID ~ changes the target ODT set.
The ODT set is done on a lookup table based on many factors. Many i don't get yet due to complete lack of information on this part.
Soo to notice you have ODT issues
You will have to isolate all other issues, to know its an "inter-cpu" issue and not a voltage communication issue.
Voltage (point) you set doesnt matter that much. Voltage alone doesnt mean anything.
Its the (A) you get out, based on the ODT set. Which on Intel's side is done automatically and passively automatically "side influence due to X , or side issue".
EDIT:
Basically if you touch SA
You shift ODT and so you shift the strength of VDD2_CPU and VDDQ_CPU.
The shifting/correction is done by the Bios itself.
Hence the "side issue" outcome of touching SA is ~ high SA , lower VDDQ<->VDDQ delta.
High SA, low delta and high voltage requirements. In some cases even higher CPU_AUX supply to feed this practice.
I wish we would have exact access to all ODTs , instead of having to hope the Board corrects this well.
But it is what it is. Here on Intel we work with imaginable delta's and learn from the Board/Bios behavior.
At least we the users. I'd like to be in the Dev position and take over control of the Bios myself~~
Currently assembled, if it behaves well, I don't want to reassemble or build it.
With a small brush that I cut short, I carefully cleaned it with 96% alcohol available in pharmacies.
The cooling water is also pharmacy distilled water, not from the gas station. It also has little alcohol.
Unfortunately, when I bought the motherboard, the previous owner left thermal paste on the socket contacts.
I think I didn't clean it thoroughly the first time, and I was also afraid of causing damage.
That's why I cut the brush short, because I only spread it with it first.
My eyes are bad, I took a lot of photos, then I saw that there was a strand of hair. Great...Pic
I managed to get it out of there too.
I don't want to lull myself into dreams, but the way I put it together, it immediately worked as it should have long ago.
I will try the 8200 profile soon, 4 rounds of TM5 have already been run, but I didn't have more time.
Of course, it could also be that I was lucky, but I hope it remains stable.
Cold water has nothing to do with it, as it was in a cold water cycle all December. I temporarily tested it in January and February with normal, 25-30°C water.
We'll see, I don't want to make a hasty statement, I'm confident that we'll still be able to move forward.
It is being said that the G4 XOC bioses have a changed V/F curve and there is a warning what to use and which switch not to use.
I dont understand the warning, and dont trust the bios.
Given it uses 1002 as a base, and even 0081 is a good Gear4 overclocker ~ maybe don't risk your CPU life.
It will not gift you anything except potential Gear2 trouble. Our higher clock is hard work, not a bios issue.
I think I didn't clean it thoroughly the first time, and I was also afraid of causing damage.
That's why I cut the brush short, because I only spread it with it first.
My eyes are bad, I took a lot of photos, then I saw that there was a strand of hair. Great...Pic
I managed to get it out of there too.
It was very hard to spot, as you simply couldn't see it in many photos. The angle of the light showed it magnified in one of the pictures. Anyway, it's not there anymore, the paste is washed out.
Very small, thin, it was impossible to notice. It was by chance that I saw him.
However, the lga1700 is very sensitive. I have never read so many problems as with this socket.
OK, dirt is another topic.
Gestern Abend kurz versucht 8200C36.
SA geblieben 1.125V
MC anstatt 1.3500V bis 1.37500V (1.38750 auch probiert)
TX/VDD/VDDQ:
1.21/1.47/1.41V Bootet, aber Y 2.5b "Coffe" error
1.24/1.53/1.44V 20. Mins error
1.25/1.53/1.47V komisch error. Drei virtel Stunden später drei error, danach lief wieder ohne error.
544 trfc kann sein, das es zu scharf ist.
1.47V über 10 Cycle..., finde ich es zu hoch.
Es wäre schön 1.44V VDDQ max.
OK, es war schnell, ich wollte 8200 ausprobieren.
Oder vielleich besser 8400C38 probieren?
Am Wochenende werde ich es weitermachen.
Ich bin schon gespannt
Bilder aufladen will ich nicht hier, so wie so, sie sind nicht errorfrei. Wenn ich es schaffe, dann.
Scau mal moi
Edit:
Eine Frage:
Statt PC aus, ist es gut?
Änderung, F10--->Startet, durch boot Retry Button.
Oder besser stomlos, ohne Strom einschalten, kein LED, nichs, dann PSU an, und starten?
geizhals.at
