Viper steel 3600 low latency on XMP
I bought the RAM I thought it would be better than the old cartridges from the dandy, but in the end, although they are higher in frequency, the results are deplorable in the Aida benchmark
I do not understand why after setting the timings from the calculator it does not work for me, the chips seem to be normal
you can ask - why do you need the results in the Aida benchmark? If it's just numbers - hack his artmani, prescribe yourself 500 thousand mb / s read / write and get high.
The disastrous result of data recording, RAM in the scrap. On ddr3 read write copy go flush.
Sweety_Mustard
In the gaming benchmark, too, everything is sadly old memory for 2400 overclocked to 3200 gives more fps than viper steel at 3600
The S
timings are too high for this kind of memory. In general, these are normal indicators, especially for this motherboard (assrock - translated as pebbles on the butt).
Sweety_Mustard
The same sore with data recording, disorder.
Spoilerhttps: //dtf.ru/ask/268512-nizkoe-znachenie-zapisi-v-benchmarke-pamyati-aida64
Solution
Spoiler prokommentirovala-raznuyu-propusknuyu-sposobnost-ddr4-protsessorov-ryzen-7-3700x-i-ryzen-9-3900x.html
Marsj
will be filled with all the owners of 3xxxx and will complain about the low write speed XD
- You know, I want to buy a computer - and what exactly will be required of it - mb high fps in games, fast video processing?
- No, I don’t need it - I want a high write speed to memory!
Sweety_Mustard
Yes, what does the write speed have to do with it, I talked about latency and I repeat once again in games the situation with FPS is such that there is practically no difference from the RAM, which is 2 times cheaper, the timings are not set lower even with 1.4 trips, all the juices were squeezed out of these dies hmm
The S
In general, be glad that you managed to get stable 3600 on your deshman 4-slot motherboard with the official limit of 3200. Besides, your kit "from the factory" does not have very good timings. Below I will throw my own, too, not the best 3400, but I think the difference is obvious (copying sags due to the peculiarities of hynix, everything would be smooth on Samsung).
Experiment with the calculator, choose other frequencies, for example. Start with a preset where the main timings do not change, but only the secondary ones.
Marsj
What are you? We are happy for you and your old stuff .. Now go study the features of Zen 2 and you will understand that this is how it should be on this architecture and such a drawdown will be compensated.
Deshang
Yes, I'm trying to experiment, but the timings cannot be reduced at all, and I changed the secondary ones using the calculator, but the effect is minimal
The S
The S wrote:
I don’t understand why after setting the timings from the calculator it doesn’t work for me, the chips seem to be normal
Because the numbers from the calculator are needed, first of all, as a guideline, and then everything depends on the "success" of the hardware and knowledge / skills of the one who accelerates. Well, of course, you need a lot of time if you want to squeeze out the maximum; the "insert and forget" option will not work.
By the way, I advise you to look at the motherboard manufacturer's website - there is a BIOS with a newer version of AGESA - this also affects overclocking. Of course, it's worth taking a look at the overclocking topic on memory raisens: it will be easier to understand what and how to do, or that you shouldn't do it))
The S
Well, I repeat, the limitation on the part of the motherboard is also very influential and you are still lucky.
Deshang
Launched at 3800 16-20-22-38, but even in the same CS, the FPS dropped almost 1.5 times compared to the old RAM
Deshang
I did not check for stability, but I played normally for two hours, then I decided to try hmr and it is very crooked constantly crashes and freezes