Reasoning about AMD and performance in older games
Hey! Based on my own observations (and not only), I noticed that on AMD hardware, especially the AMD (CPU) + Radeon (GP) bundle, many old games, for example, GTA5, Far edge 3/4, Crisis 3, behave on this gland is inadequate. Take, for example, similar bundles - ryzen 5 2600 + rx 580 and i5 9400f + gtx 1060 (system on both Windows 10 pro). Even if we equalize the frequencies in frequencies, fps and stability are higher on the intel + nvidia bundle, that is, frametime is more stable and there are no friezes, while there are "red" friezes on the assembly. Question: AMD drivers are so unoptimized in old games, or are they more guilty of percentages, perhaps weak cores ????Outlaster
He wants to say that the ryzen 1000-2000 prots are so slow that even in the stock they cannot work, so it is necessary to drive memory, underestimate the timings is directly critical, because of their controller and cache, for this it is better for those who are not experienced in this matter to go and collect intel, and experienced, in principle, too =) Why the hell to bother with Amd when intel gives the same profit, I'm talking about 2600-9400 ... Well, the maximum that can be conjured up if you have a mother B chipset is timings and that's it ... Intel purely out of the box and on the way!
Ryzen 3000 -5000 in this regard are less capricious, but also require manipulation.
Ryzen 5 1600 should be compared with i5 7400.
Ryzen 5 2600 should be compared with i5 8400.
i5 9400 and i5 10400 should be compared with ryzen 5 3600. It is
incorrect to compare Intel 9000 with the first two generations of risens.
I haven't had amd cards for a long time, but with amd processors how to say ... I had an amd fx 6100 16GB 1333 memory and a powerful vidyuha 690 is listed and all this is in low resolution at 1600x1200 1600x900, but then I switched to full cd so that load vidyuhu. and without overclocking the processor, there were always freezes. by friezes, I mean some unpleasant jerks of the micro-freeze image that are not associated with low FPS, and so the processor simply does not take out. there was also a very low performance in old games, some that are not very old (for example, Feyball 3 had a performance drop of up to 20fps, maybe I don't remember with a defeat without 690 with an active slam at 1600x1200 it's hell). then I separated the processor without touching the bus and the north bridge - the friezes were basically gone and, in general, the performance was enough for me for a cinematic FPS. advised overclocking the south bridge, I isolated it, he separated it, hell began, the performance became uneven and moreover, the time began to flow in the games incorrectly, they could slow down or accelerate and the fms could finally fall. in synthetics, the north bridge and memory gave nothing. the only benchmarks of interest to me received a boost from the complex overclocking of the bus - this is a real storm 3d, it seems, and a startales benchmark. The latter, in my opinion, should pull the GPU with a high FPS on the cps. and the taut, as it were, the processor was worse than the old overclocked intel about performance per core. then, in short, I decided to replace the vidyuhu with 2060s. if the performance suffered due to the fact that the 690 had some kind of plx chip and it had to split the express 2.0, 5.0GB sec in half, then with one ordinary card the performance of the stone did not change in any way. after how the motherboard deteriorated, I replaced it and rolled it on an unknown fx 6100 in games from 2060s - an absolutely identical situation freezes, lags, glitches and other whistles. so there was nothing to do with the slide. And moreover, if you put the FPS lock below the one on which the percent drags the friezes go. I also lost in cybcrpuink 2077 on ultras on amd fx - it seems to work fine when you sit down on a motorcycle there the FPS drops sharply for some reason, but there is no processor in the minimum system requirements.
I have thoughts on this in relation to my processor - this means its scheduler and turbo boost. I also had problems that in games where 690 did not have enough memory, there could be drakefaces and freezes. but that's a slightly different story.
I did not have any ruzen, but I carefully watched some video tests (for 16 streaming or more) and gaming test data and came to the conclusion that these are good processors, but some described free prizes. recently, someone seemed to describe that he had a witcher 3 for 2700 or what such a frieze gives.
but there are friezes and there are few nuclear i3 and i5 when they hammer in a high fps on a sotochku. and actually my guess is that all processors might be hard to find. it will have to overclock the games up to 300-500fps. and looking for that the openers were dancing the video card over which I trolled and joked about what they turned out to be right.
Therefore, in short, my opinion is that if there are friezes, you need to lock FPS 30 programmatically so that the game does not overclock more than 30. if the friezes do not go away, the problem is deeper in something. And if they leave, then the problem is solved, in general, then 4k 30fps, as it were, is more and not really dando.
Regarding video card drivers in general, it is rather vague there. In my opinion, the friezes from the card are mainly due to lack of memory and power limit and trolling. but even with the latter, in theory, there should be a tangible drop in performance (although I'm not sure for how long the frequencies are reduced and how the turbo boost works for video cards.) For fidelity, you can put the card at constant frequencies - I'm not sure if this can be done with modern software - the essence of the card will always yell and warm up like a load, because it will be prohibited from dropping frequencies.
The low performance of the card is in its own specificity, but I think an experienced PeKa gamer will feel the difference with his spinal cord if his card stably rolls over 20-30 fps or out of the blue 40-50 fps gives some dips and pauses.
Outlaster
For its time it was the norm, but it is already beginning to become obsolete.
But this does not prevent him in our time in a modern game like Cyberpunk 2077 from raping his direct competitor into all the cracks
https://overclockers.ru/lab/show/107809/cyberpunk-2077-testirovanie-chetyreh-pokolenij-processorov
Edgarrr
is it in stock or overclocking? and with what kind of test? and still 8400/9400 Intel comes out better here ...
Outlaster
Everything is indicated in the test on the link.
Regarding better not better, I have already written above what to compare with.
And you again have incompetent comparisons, respectively, incompetent conclusions.
Outlaster
You have nothing to justify, the first ruzen products were designed for enthusiasts and the displacement of a competitor from the market, those who bought 1600 Ryazan knew that it had to be driven and driven with memory.
I inserted and forgot how the intol is already 3xxx-5xxx ruzen.
If you don't know why you need a product, then there is nothing to buy it.
I bought a tractor, and you want the speed of a Ferrari.
Lex-one
So the worse the graph and the older the game, the more demanding the process? Is this surrealism?
Outlaster
No, the fewer cores it can load, therefore the FPS is weaker.
Take even GTA 4.
Edgarr
Lex-one
Here the topic is probably not about performance. At the end of 2021, TS bought a system for gambling for 1600. But it turned out that 9400 is suddenly better in a single-stream. And he had a brain intel.
Patrick Star Online
So, as far as I remember, the percentage is two times cheaper than the same 9400.
I would take a 10100f for myself, and I wouldn’t fool anyone.
Whoever collects for 1600 knows what is going on and looks at tests / reviews before buying.
Outlaster
r5 1600 + gtx 1070 + 16 gb of RAM at 3200 + ssd = these old games are fine. I would not say that there are some microfreezes, yes, the FPS at high is not cosmic, but the hardware is penny.