2311
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 23:05 »
Cyberpunk 2077 nu suporta (inca) AMD FidelityFX CAS, deci nu e nimic de comparat cu DLSS.
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
2312
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 22:39 »
Asta e. Unii spera ca leak-urile/speculatiile sunt reale, altii spera sa nu fie reale.
.....................
Uite aici leak, cand nici nu s-a anutat nimic despre existenta lor.
.....................
Uite aici leak, cand nici nu s-a anutat nimic despre existenta lor.
2313
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 21:15 »
Lumea nu stie care sunt frecventele de referinta (oficial) decat tot din leak-uri, dar cei care dau leak-uri vor sa para cat mai obiectivi, altfel se pierde increderea in ei. Din moment ce au testat placile (presupunand ca e reala afirmatia) inseamna ca au si frecventele de referinta la placile AMD, iar la Nvidia sunt cunoscute.
AMD Radeon RX 6800 XT specs ("Based off of speculation/reddit and other sources online.")
AMD Radeon RX 6800 XT specs ("Based off of speculation/reddit and other sources online.")
2314
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 20:53 »
Nu fi obraznic ca nu eu le-am facut si sunt publicate de un sit respectabil, cu mentiunea de "zvon", pentru ca nu e nimic oficial. Iar comparatiile se fac la frecventele de referinta ale placilor comparate, in faza asta.
2315
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 20:48 »
E foarte suspect faptul ca nimeni nu vrea sa spuna nimic despre review-uri, cu 2 zile inainte de "dezvaluire". E posibil ca miercuri sa nu aflam nimic referitor la adevarata performanta si pretul real, ci doar povesti... Adica e secret si momentul in care se ridica secretul.
Daca ar fi sigur ca NDA se ridica poimaine, ar fi clar ca vor fi si la raft si la ce pret. Daca o freaca cu NDA-ul inca vreo 2-3 saptamani, nu cred ca vom afla nimic din ceea ce ne intereseaza, in 28.
Daca ar fi sigur ca NDA se ridica poimaine, ar fi clar ca vor fi si la raft si la ce pret. Daca o freaca cu NDA-ul inca vreo 2-3 saptamani, nu cred ca vom afla nimic din ceea ce ne intereseaza, in 28.
2316
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 20:17 »Hmm, arata bine, poate fac schimbarea de la 3080 la 6900XTDaca 6900XT bate 3090, probabil o sa coste 1500$ (MSRP), 2000$ in primele cateva luni, si o sa consume cat un calorifer. Sa vedem daca faci schimbul in conditiile astea...
2317
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 19:48 »
Doar ca rezultatul respectiv nu e cu placa la frecventele de referinta, e foarte OC, deci e irelevant.
2318
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 19:38 »
Am mutat postul in threadul potrivit.
2319
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 19:35 »
Alleged 3DMark scores of AMD Radeon RX 6800XT
In mai multe teste, 6800XT se pozitioneaza la jumatatea distantei (ca performanta) intre 3080 si 3090. E de asteptat ca 6800 se se pozitioneze tot la jumatate intre 2080 Ti si 3080 (in rasterizare normala). Se anunta o realizare deosebita din partea AMD.
............................................
Solutia AMD similara cu DLSS e FidelityFX Upscaling, care da rezultate foarte bune, uneori chiar mai bune decat DLSS, mai ales in optiunea pentru calitate.
"Yet astoundingly, AMD's FidelityFX CAS, which is platform agnostic, wins out against the DLSS "quality" setting."
In mai multe teste, 6800XT se pozitioneaza la jumatatea distantei (ca performanta) intre 3080 si 3090. E de asteptat ca 6800 se se pozitioneze tot la jumatate intre 2080 Ti si 3080 (in rasterizare normala). Se anunta o realizare deosebita din partea AMD.
............................................
Solutia AMD similara cu DLSS e FidelityFX Upscaling, care da rezultate foarte bune, uneori chiar mai bune decat DLSS, mai ales in optiunea pentru calitate.
"Yet astoundingly, AMD's FidelityFX CAS, which is platform agnostic, wins out against the DLSS "quality" setting."
2320
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 17:37 »
V-am spus asta de vreo saptamana (postul pare a fi disparut la curatenie). Adica o placa pe cipul GA102 castrat mai mult decat pentru 3080, in loc de GA104 full. Are sens pentru ca sunt mai multe cipuri GA102 partial functionale decat GA104 perfect/full functionale. Pe GA102 pot pune 10GB VRAM, pe GA 104 ar fi trebuit sa mareasca memoria la 16GB sau sa o mentina la 8GB.
Aspectul derutant il constituie consumul urias anticipat.
Aspectul derutant il constituie consumul urias anticipat.
2321
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 16:19 »
Au scos si DLSS 2.1.
The key features of DLSS 2.1
Excellent Image Quality: DLSS 2.1 offers superior image quality that is comparable to the normal, native resolution while rendering only one quarter to one half of the pixels. It utilises new temporal feedback techniques for sharper image clarity and improved stability from frame to frame.
Great Scaling Across For All RTX GPUs and Resolutions: Nvidia’s new AI network is more efficient at utilising Tensor Cores to execute at double the speed of Nvidia’s original. This vastly improves frame rates and eliminates previous limitations on which graphics cards, settings, and resolutions could be enabled.
All Games On One Network: The original DLSS required training the AI network for each new game. DLSS 2.1 is different, in that it trains using non-game-specific content, distributing a generalised network that works across DLSS supported games. This means faster game response times, and ultimately the launch of more DLSS games in the future.
Customisable Image Options: The real beauty of DLSS 2.1 is that it offers users 3 image quality modes; Quality, Balanced and Performance. These modes essentially control the game’s internal rendering resolution. With Performance mode it enables up to 4x super resolution - so from 1080p up to 4K.
E foarte tare treaba cu DLSS si se indreapta spre solutia universala oferita de cipurile specializate cu AI care fac scalarea in TV-urile 8k! ("...it trains AI using non-game-specific content...")
.........................................................................
Solutia Samsung cu AI pentru reconstructie de imagine la 8k.
The key features of DLSS 2.1
Excellent Image Quality: DLSS 2.1 offers superior image quality that is comparable to the normal, native resolution while rendering only one quarter to one half of the pixels. It utilises new temporal feedback techniques for sharper image clarity and improved stability from frame to frame.
Great Scaling Across For All RTX GPUs and Resolutions: Nvidia’s new AI network is more efficient at utilising Tensor Cores to execute at double the speed of Nvidia’s original. This vastly improves frame rates and eliminates previous limitations on which graphics cards, settings, and resolutions could be enabled.
All Games On One Network: The original DLSS required training the AI network for each new game. DLSS 2.1 is different, in that it trains using non-game-specific content, distributing a generalised network that works across DLSS supported games. This means faster game response times, and ultimately the launch of more DLSS games in the future.
Customisable Image Options: The real beauty of DLSS 2.1 is that it offers users 3 image quality modes; Quality, Balanced and Performance. These modes essentially control the game’s internal rendering resolution. With Performance mode it enables up to 4x super resolution - so from 1080p up to 4K.
E foarte tare treaba cu DLSS si se indreapta spre solutia universala oferita de cipurile specializate cu AI care fac scalarea in TV-urile 8k! ("...it trains AI using non-game-specific content...")
.........................................................................
Solutia Samsung cu AI pentru reconstructie de imagine la 8k.
2322
Hardware / NVIDIA next-gen Ampere GPU on 7nm
« on: 26 October 2020, 13:30 »
DLSS 2.0 aduce performanta enorma, si peste 100% in unele situatii. Deocamdata se limiteaza la putine jocuri, dar nu peste mult timp va deveni "universal", respectiv va putea actiona in orice joc, pentru ca Nvidia schimba algoritmul care antreneaza AI-ul si va renunta la necesitatea de a folosi neaparat texturi specifice pentru antrenament.
2323
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 03:23 »
"Nvidia have a habit lately of selling cards with Vram that falls short of the full potential.
The problem gets worse over time and results in you using lower settings than the card can handle (you may be stuck at 60FPS, but have to keep settings lower to avoid stuttering).
Just bare it in mind."
E normal ca la lansare totul e calculat astfel incat placile sa se descurce cu ceea ce este lansat pe piata. Cand apar problemele se apeleaza la upgrade si se rezolva din nou.
........................................
Problema a fost sesizata si bine evidentiata cu multi ani in urma, cand s-a lansat Fury X.
"Rise of the Tomb Raider is the most high profile example.
When you use the highest texture setting a warning is flashed on screen that tells you this setting will need more than 4 gb's of vram.
If you use that setting with a Fury (4gb's of HBM) you experience problems with how it runs, even though the fps is still good it stutters, sometimes badly, This is at 1080p.
Turning the texture setting down one level solves the problem."
Degeaba au bagat memorie rapida HBM, prea putina memorie i-a dus in cap. Unii au prins ideea, altii nu. Se rezolva cu reducerea nivelului de calitate, dar ca idee.
The problem gets worse over time and results in you using lower settings than the card can handle (you may be stuck at 60FPS, but have to keep settings lower to avoid stuttering).
Just bare it in mind."
E normal ca la lansare totul e calculat astfel incat placile sa se descurce cu ceea ce este lansat pe piata. Cand apar problemele se apeleaza la upgrade si se rezolva din nou.
........................................
Problema a fost sesizata si bine evidentiata cu multi ani in urma, cand s-a lansat Fury X.
"Rise of the Tomb Raider is the most high profile example.
When you use the highest texture setting a warning is flashed on screen that tells you this setting will need more than 4 gb's of vram.
If you use that setting with a Fury (4gb's of HBM) you experience problems with how it runs, even though the fps is still good it stutters, sometimes badly, This is at 1080p.
Turning the texture setting down one level solves the problem."
Degeaba au bagat memorie rapida HBM, prea putina memorie i-a dus in cap. Unii au prins ideea, altii nu. Se rezolva cu reducerea nivelului de calitate, dar ca idee.
2324
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 26 October 2020, 01:24 »
Nu pricepeti o treaba simpla. Aplicatia incarca in memorie atata informatie cat considera ca are nevoie (indiferent daca e folosita sau nu). Daca cipul grafic e slab, performanta e trasa in jos de cipul grafic. Problemele apar daca cipul grafic e puternic si memoria e insuficienta, nu invers.
Ce spune chiar Nvidia:
"There are a number of hardware monitoring tools and utilities that claim to report GPU Memory Usage. Note that these tools do not report actual GPU memory usage. Instead, they report memory allocation, the amount of memory requested by the application.
This number can vary for a variety of reasons, and should not be used as an indicator of the amount of GPU memory that the application is actually using. In fact, in some cases the application may request all of the GPU’s memory, whether it actually needs it or not. To illustrate this issue, we have included memory allocation logs running Tom Clancy’s The Division 2. In this example running on a GeForce RTX 2080 SUPER (8GB) and RTX 2080 Ti (11GB) using the same settings (4K with the Ultra preset), the game allocates almost all of each GPU’s framebuffer:
RTX 2080 SUPER (8GB) RTX 2080 Ti (11GB)
AVG: 41 AVG 49
MEM 7393 MB MEM 10531 MB
As you can see, both GPUs are running at high frames (41 fps vs 49 fps, with both GPUs running smoothly), with the game allocating as much memory as it can, which resulted in memory consumption of 7.3 GB on the RTX 2080 Super and 10.5 GB on the RTX 2080 Ti."
In principiu tocmai s-au dat de gol, pentru ca, consumul de memorie creste impreuna cu performanta cipului grafic. Acum o sa avem RTX 3070 cu performanta de 2080 Ti, care 2080 Ti are nevoie de 10,5GB de memorie pentru 49fps, iar RTX 3070 va avea doar 8GB memorie disponibila ca sa scoata tot in jur de 49fps. Urmeaza sa vedem ce rezulta in review-uri din asta.
E valabil si pentru AMD, nu degeaba s-au decis sa ofere 16GB VRAM pentru placile din clasa superioara.
Ce spune chiar Nvidia:
"There are a number of hardware monitoring tools and utilities that claim to report GPU Memory Usage. Note that these tools do not report actual GPU memory usage. Instead, they report memory allocation, the amount of memory requested by the application.
This number can vary for a variety of reasons, and should not be used as an indicator of the amount of GPU memory that the application is actually using. In fact, in some cases the application may request all of the GPU’s memory, whether it actually needs it or not. To illustrate this issue, we have included memory allocation logs running Tom Clancy’s The Division 2. In this example running on a GeForce RTX 2080 SUPER (8GB) and RTX 2080 Ti (11GB) using the same settings (4K with the Ultra preset), the game allocates almost all of each GPU’s framebuffer:
RTX 2080 SUPER (8GB) RTX 2080 Ti (11GB)
AVG: 41 AVG 49
MEM 7393 MB MEM 10531 MB
As you can see, both GPUs are running at high frames (41 fps vs 49 fps, with both GPUs running smoothly), with the game allocating as much memory as it can, which resulted in memory consumption of 7.3 GB on the RTX 2080 Super and 10.5 GB on the RTX 2080 Ti."
In principiu tocmai s-au dat de gol, pentru ca, consumul de memorie creste impreuna cu performanta cipului grafic. Acum o sa avem RTX 3070 cu performanta de 2080 Ti, care 2080 Ti are nevoie de 10,5GB de memorie pentru 49fps, iar RTX 3070 va avea doar 8GB memorie disponibila ca sa scoata tot in jur de 49fps. Urmeaza sa vedem ce rezulta in review-uri din asta.
E valabil si pentru AMD, nu degeaba s-au decis sa ofere 16GB VRAM pentru placile din clasa superioara.
2325
Hardware / AMD Navi GPU & Nextgen Architecture
« on: 25 October 2020, 23:25 »Daca faci o corelare cu performanta, Radeon VII nu este ajutat defel de cei 16GB...Daca faci o corelare cu performanta, ar trebui sa-ti dai seama repede ca un fps mai mare va cere si mai mult vram, adica Radeon VII consuma peste 11GB VRAM la un fps mic.