Built-in graphics card.

MunchkiN 616
hard modes say it was 10-15фпс was the 20-30фпс
I threw the screen was 25 became 50fps ! In this case, the 280x is constantly low fps and here especially in Hitman.
the difference in frequency mostly. and I have a AMD raden conditional
But this nonsense you're writing and otmazyvaysya not from words ! By your logic RX5500XT should be weaker than the 280x because the tire is less and memory speed ! 5500XT just faster not even two and three times )))
Your logic is lame (((
Because they use different GPU and architecture of the L2 cache as it suggests, cannot be divorced from the graphics processor to compare memory and bus !
By your logic, AMD just renames your old GPU !? This is crazy !
All just uncle don't stick it to me here for example;
Nvidia has 1080TI CACHE L2 2.5 mb
Nvidia 2080TI has the L2 CACHE 5.5 mb
Same thing from Nvidia just the difference you will see at the end of life 1080TI when it will be drained due to the old architecture.
дх12 I do not take into account
Well, it's apparent schizophrenia as RTX, and so only runs on Dx12 and Vulkan well, still Metal but certainly not in DX11.

MunchkiN 616
Yes not prove his numbers are not numbers, proof, not proof. You know, as always with him.

X_ray_83 wrote:
Because they use different GPU and architecture of the L2 cache as it suggests, cannot be divorced from the graphics processor to compare memory and bus !
By your logic, AMD just renames your old GPU !? This is crazy !
in General, I tend to pay attention to the number of Shader processors. in the second place the brine and texture to the blocks then the frequency then memory. the bus as such I'm not interested in 256 or 512 bits there. it plays a role mainly at high resolutions and high fps.
about the cache graphics card do I have to peck my head is empty because I don't know how it is used and what affects.
about the fact that AMD makes the old chip and glue on a new label - I don't think so. most likely there is some kind rebilding optimization and redesign just like it was with the BH 4870 and 4890, or something there. it's kind of the same but the layout of the chip is different from what there frequencies were higher. but mathematically it's the same chip and magic when I was one architecture and became another, and it is mathematically significantly faster at the same frequencies and compute blocks is very likely is rarely.
X_ray_83 wrote:
I threw the screen was 25 became 50fps !
have GameSpy like as so signified minimum and maximum fps. on the screen appears рх380х-570 34 44 against the lower limit of measurement of fps. 40 to 53, the upper bound of intended fps. if diagonal view 34 and 53 the increase is certainly impressive, peck-peck

...as RTX, and so only runs on Dx12 and Vulkan
well, the Radeon is no p does not work as far as I know so that sense to consider it. even дх12 and the volcano, in theory, can speed up the selection of requests on rendering. but then there are a couple of benchmarks and games, this yet and not very soon will appear the most likely. and дх12 what I visually associate with rthcom and all such as dh11 with tessellations, DH10 with a rich depth buffer and multitexture etc

Lex-one
If you have a processor running in isolation from memory I've nothing new to tell
we all talked about that in 2002 that the main CACHE is always in any device depends on the speed and so on.
Although you probably don't even know the name of Nvidia in the 90s that you can tell me !

MunchkiN 616
in General, I tend to pay attention to the amount of Shader processors
Again this is a mistake and are rude, have RX570 better than RX5500XT !
In fact RX570 merges new 5500XT
Here:
RX570

RX5500XT

Why RX570 merges Shader not helped !? It's very simple the first CACHE L1 128kb and not more 16kb !
This is stupid other percent other architecture and this is because it is faster shaders and stuff there to the one place GPU is stupidly fast for the beat.
have GameSpy like as so signified minimum and maximum fps
I learned they denote the minimum and average, and since the 280X is 3gb memory no ultra out of the question maybe it's friezes and the pain in the same Hitman 2. Subtract 10 fps at once.
You Radeona do not understand may not be worth it to try !?
well, the Radeon is no p does not work as far as I know so that sense to consider it.
The point is that there are games which are purely for Dx12 ! So that excuse is so-so.
On PS5 works there as Vulkan or OpenGL, but not the essence Dx11 is the past.

X_ray_83
have рх5500 frequency is higher. it compensates for the wider Shader broom high frequency Makhach.
the essence of my understanding of architecture is a specific job at a certain frequency equal to the different agents and their difference in performance. for example, if we take processors i7 2600 to i7 7700 (which is the last and7 according to the scheme 4*8) transferring them to a single frequency (e.g. 3 GHz) and send it to do benchmarks. some scatter but there will be between 6700 and 7700 I think it will be very minor. this effect is there any optimization of decisions taken in the architecture. or take a 4 GHz AMD FKH and its counterpart ruzena by the number of threads at the same frequency - I think there will be much difference.
the same logic works with the GPU and as I wrote above, there are significant non-obvious transitions that busstat performance in common applications and has extensive growth. and all 3-digit cards from AMD + 7ххх generation just Pitch in my head fit in the extensive niche performance that tale.
it is clear that there can be popley to count and there are 2 ways to raise them is to do more parallel processors, or to increase the frequency (and increase in frequency within certain limits is often more effective at times in common tasks) and most likely, this RKH 5500XT from this point of view is so predictable Zhu neatly fit into some pattern of performance (budget card especially from AMD did not interest me so crazy I actually laziness) it can be assumed that she should have to jostle with the level of performance of 2015, i.e. with the line of any 390х 480 580 etc and there may even by 10-15 per cent they will overtake.

MunchkiN 616
have рх5500 frequency is higher. it compensates for the wider Shader broom high frequency Makhach
And now you have to measure performance frequency ! You're just a walking stereotype.
transfer them to a single frequency (e.g. 3 GHz)
Intel has not changed the architecture since 2008, it is still a Core2Duo actually ! The comparison is stupid all they did manipulyatsii with the ring bus and the rest is not a strong architectural changes.
Here, the growth is 50% relative to the old architecture and not a measly 20-30%.
Stupid comparison accounts.
the same logic works with the GPU and as I wrote above, there are significant non-obvious transitions
It does not work with GPU, by your logic RX5500XT shit and which merges RX570 ! In reality, this is not because Navi 7nm are also Packed more densely on the count of transistors in contrast to 14нм.
Even on the count of transistors 6 million vs 5 million RX570 merges with a larger physical size !
You to me here about what frequency mold.... this is a completely different GPU and different number of transistors, you can't just pack in 7nm old architecture of Polaris.
TA RKH 5500XT from this point of view Zhu neatly predictable fit
The pattern is simple cutbacks RX5700XT because it's so fast ! The problem is that we have not seen the Top and not what a start from word at all.
line any 390х 480 580
15% is your guess no more at 1080P the difference is sometimes 30% !

MunchkiN 616
In General, to compare the frequency of different architecture is the idiocy of the 90's !
You still RISC percents compare c X86 frequency one can be )))

MunchkiN 616
And how do you decided that the performance on the beat the same for Navi and Polaris !?
He invented or decided to fantasize about the payment frequency" and other nonsense !?
Tin and these people suggest the graphics card, measuring performance frequency ))) Ahaha )))

X_ray_83
if the card is structurally the same number of elements - appropriate direct comparison. a direct comparison between, for example 290 and 390 in my opinion appropriate more than. and 280х from 570 if to speak about the performance of the chip.
you can also talk about conditional performance impact for me is the top card in a certain year and when we go out madly and lo ENDA I say something like гтх1060 map for 2015 and nothing controversial that is not see because of the dispersion performance 980-980ти.

and Yes. that is not so much the frequency but the theoretical math performance it is possible to measure such cards to work in hard modes. however, AMD with NVIDIA's impossible to measure directly because traditionally at a higher Mat performance AMD less effective usually was. but as predictive ability in the framework of some cards looking at the claimed features can well imagine the performance of the card. of course sometimes it happens that it is not so and there are many factors.
what the actual conversation? or to measure the amount of cache and its performance with a cpu is generally much does not work or is very sluggish. I have had q 9650 or couple there that I don't like him don't Smok squeeze acceleration of the tire i5 2500 but I think the cache is better and in 3D Marche plums. so there is probably architecture....
what conversation do not understand what's the transistors.

MunchkiN 616
and 280х from 570 if to speak about the performance of the chip.
Another delusion 280x is old the process is, structurally, there's different things you can't just so packaging in 28nm 14нм ! For this and created a new architecture.
You compare the incomparable Tesla and Lada.
the declared characteristics can well imagine the performance of the card.
It is not possible because we have a lot of cores in the graphics card and not 2-6 as the processor ! So stop talking nonsense.
what the actual conversation?
The talk is that under the new process creates a new architecture and not otherwise, you're all the same arhitekturnye features doesn't matter to you you and confirmed with Dx12 and Vulkan !
I'm not talking about HEVC hardware-based codec that is supported RX570 and a resolution of 8k which can only dream of 280x.
what conversation do not understand what's the transistors.
Well, actually it's brutal (((( Moore's Law read it for the GPU too common !