Search results

C
Can more CPU power help when GPU is at 100%?

With AMD card thats possible in DX11 CPU bound cases since AMD GPU load reader apparently works in strange ways. What i mean is than even when AMD GPU load showing at 100% usage in some rare cases that dont mean the GPU is really at 100% because with faster CPU is possible to extract some extra...
- chummy
- Post #14
- Jan 30, 2017
- Forum: Graphics Cards
C
Awful AMD VCE performance

Staxrip with NVIDIA H264 encoding, choose some compressed file up to 50Mb bitrate to avoid disk bottleneck than can occur with lossless source and slow down encoding.
- chummy
- Post #94
- Dec 17, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

While jet flying the blocky artifacts nvenc/shadowplay produces in the sky is bad and VCE AVC look better there. But in the last run the quality from nvenc is better than VCE AVC. AMD HEVC cause too much ringing/mosquito noise at edges in high motion, i dont like how it looks much but it has...
- chummy
- Post #92
- Dec 16, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

I only can see the Relive AVc is duplicating a lot of frames while Relive HEVC has lower duplicating and GameDVR is free of duplicating. The cause i dont know, i see Xaymar saying than Relive records with variable framerate but need to check if its fake VFR like Shadowplay/GameDVR or if is a...
- chummy
- Post #90
- Dec 11, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

I uploaded a video comparison sometime ago for BF4 scene and for a scene from a movie to check quality difference. I used 4Mbit for recording waiting Youtube H264 to reencode at this same bitrate, problem than youtube dont always reencode at 4Mbit, if the video complexity is low it will use...
- chummy
- Post #87
- Dec 11, 2016
- Forum: Graphics Cards
C
AMD ReLive

Always record to dedicated HDD to avoid problems. But if single disk I/O is fast enough for game read+record writes and it dont overload disk i dont know if too much performance can be dropped, something to find out.
- chummy
- Post #30
- Dec 8, 2016
- Forum: AMD
C
AMD ReLive

The performance impact is very good. On my test with bf4 Win10DVR achieved lowest performance loss with 3,4% to NVENC1 and 4,4% to VCE2. Take notice than the performance hit vary from game to game like the anandtech test demonstrate because each game load CPU, RAM bandwidth, PCI-e and even GPU...
- chummy
- Post #28
- Dec 8, 2016
- Forum: AMD
C
Awful AMD VCE performance

@Shivansps I waiting you test encoding with VCEenc to confirm performance of Xaymar test. Xaymar test is very strange and strange yet. He messed up 290x and 390 test, he used B-frames for 390 and no b-frames for 290x and results became different but both cards have same VCE version. I have lot...
- chummy
- Post #65
- Dec 4, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Yes, i say to you test it, Rigaya(developer name). It's called VCEEnc. If you dont like messing with CL, you can download Staxrip which has VCEEnc implemented and you can encode things there, but its using old SDK. Try it out to see how VCE4 performs there. Rigaya developed NVEEnc, QSVEnc and...
- chummy
- Post #60
- Dec 2, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Lets see if now with HEVC releasing some software royalty-free it will change something, but you know Google and all big boys already created last year the Alliance for Open Media where they're developing a new codec to be released 1H 2017. I dont know if Google will adopt HEVC in this stage and...
- chummy
- Post #55
- Dec 1, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

@Shivansps do you tried Xaymar encoding test? Are it only loading encoder or put stress in 3D GPU at same time? Can you try solely encoding with Rigaya VCE encoder to see how it performs there, it use old SDK but at least with VCE 2.0 was very stable and shows true performance. Afterburner has a...
- chummy
- Post #54
- Dec 1, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

First about your last question you can send AMD a report, but Xaymar already has constant contact with AMD guy in AMF repository where he keep reporting bugs. So i guess he already has being warned about it and if there has some fix will appear in future releases. But yet users can report bugs...
- chummy
- Post #52
- Nov 29, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

When i start encoding it increase something 20-30W.
- chummy
- Post #50
- Nov 29, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

So the problem is because AMD Polaris has bad power efficiency then AMD for marketing purpose capped the power feeding to the card and it's causing performance loss when using encoder extra power. How smart is AMD with their false advertisement wanting to promote the card something it's not and...
- chummy
- Post #47
- Nov 28, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

In their VCE plugin repository has Wiki for each VCE generation and now has performance benchmark for every resolution and quality presets.The amount of performance loss from VCE 3.4 compared to 3.0 is like 50%. I dont know what the problem with AMD for doing this.. I just hope it's driver...
- chummy
- Post #42
- Nov 28, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

I seeing Xaymar added some benchmarks for H264 VCE generation and i need to laugh. VCE 3.4(2016) has slower performance than NVENC 1.0 from 2012 and even lower quality. The worse is than VCE 3.4 is slower than VCE 3.0, AMD doing downgrade over generations, this is the first time some brand...
- chummy
- Post #40
- Nov 27, 2016
- Forum: Graphics Cards
C
YouTube now officially supports up to 10-bit HDR 8K 60fps

H264 1080@30fps= ~4Mbit/s H264 1080@60fps= ~5.5Mbit/s VP9 1080@60fps= ~3.5Mbit/s VP9 1080@30= ~2.5Mbit/s H264 1440@30fps= 8-9Mbit/s The bitrate is focused in user connection bandwidth average, not in keeping your video quality. If average connection speed from users was higher then Youtube...
- chummy
- Post #11
- Nov 8, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Shadowplay use some hard coded CFR to make a constant framerate video output, i dont know exactly what they do, but they make the video to become CFR even if mediainfo read it like VFR. The way of capturing the frames make no difference how it will be encoded at framerate. Win10DVR is doing...
- chummy
- Post #32
- Nov 6, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Win10 DVR has better frame capture than OBS(far less overhead) because apparently it's using similar method to NVFBC, it feed frame to the encoder directly from GPU buffer while OBS send GPU rendered frames to system RAM first then it come back to GPU VCE encoder what cause higher latency. At...
- chummy
- Post #30
- Nov 5, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

B-frames are a bit useless in Hardware encoders. In NVEnc & Quicksync they only help making frames totally blurry when too hard to keep quality to avoid artifacting instead, remember something from HEVC/VP9. Using only P-frames make a better average frame quality for QS and NVEnc. The quality...
- chummy
- Post #24
- Nov 5, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Maybe VCE3 has some quality improvements over old one. But VCE4 from Polaris apparently loose some H264 settings like B-frames which were added only to VCE3. What Quicksync version and quality you used to compare to VCE3? Besides that VCE worked for me with old Media foundation, Xaymar...
- chummy
- Post #18
- Oct 30, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

Just to clarify, each generation of ASIC can be added quality and speed improvements, most of time they only is adding speed improvements. Quicksync1/2 SB/Ivy has no good quality like Haswell+. For VCE apparently VCE1 is the worse then VCE2,3,4 has around same H.264 quality but maybe there is...
- chummy
- Post #17
- Oct 30, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

VCE2 has lower quality than Quicksync3 in H264. VCE2 is worse than x264 Superfast while Quicksync is draw to x264 veryfast. Just tested with Witcher 3 and at same bitrate 36Mbit/s Quicksync is better than VCE, so increased VCE bitrate to 47Mbit/s but it yet lost by a few to Quicksync at 36Mbit/s...
- chummy
- Post #14
- Oct 30, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

What game quality settings was you really using in each encoder? Let's start with basics. There is something called complexity in video compression. The more detail more complex to compress, adding fast movements make it even harden to compress. In VCE video has all that shimmering, sharpness...
- chummy
- Post #10
- Oct 25, 2016
- Forum: Graphics Cards
C
What is Dynamic Memory?

Just notice than in DX12 games Dynamic RAM is being reported far higher than in DX11 in apps which read it like GPU-Z or HWinfo. Running AotS DX12 it shows 6GB VRAM usage + 5GB Dynamic RAM while the AotS process itself is consuming 7.5GB of system RAM. Running AotS in DX11 it report 6GB VRAM...
- chummy
- Post #7
- Oct 25, 2016
- Forum: Graphics Cards
C
Awful AMD VCE performance

AMD not put many effort in their VCE for years, only last month they started messing their ass for it slowly. Shadowplay record with VBR too so i dont understand your complain here. But seeing your footage yes there is a lot of blockiness in the OBS version too which keep 30Mbps and that a bit...
- chummy
- Post #3
- Oct 25, 2016
- Forum: Graphics Cards
C
Forza Horizon 3 Benchmark Thread

Yet when you stack trace APEX with process hacker you find there DX11 and DX12 functions. Game looks like a hybrid DX11/12? Other DX12 games i monitored dont call DX11 functions like APEX does. I just wondering if it makes things worse..
- chummy
- Post #109
- Sep 30, 2016
- Forum: Graphics Cards
C
Polaris vs Pascal for Video Editing/Adobe Premiere Pro CC

Premiere has stated there is no encoding GPU acceleration but the half time transcoding looks too fast for GPU effects only. CUDA and Opencl only speed up effects at all there, for sure? No gains in a pure encoding bound schema? Vegas implementation of CUDA speed up a lot the encoding giving...
- chummy
- Post #70
- Jun 26, 2016
- Forum: Graphics Cards
C
Sony Vegas GPU accelerated rendering

I use Vegas only to small changes focused more in cutting parts out and doing some split screen videos which should be encoding-bound. But in my test encoding only with a Fermi card it was capable of speed up the encoder a lot decreasing time by less than 50% with mainconcept/avc, while using...
- chummy
- Post #30
- Jan 22, 2016
- Forum: Graphics Cards
C
[Eurogamer] The state of 2GB VRAM GPUs

Bad word choosen from me, its more like some range it stick around after some time. I play 1080p, if i increase some settings like extended distance from 0 to 50%(this setting increase input lag a lot so i dont like, dont even increase much the draw distance) then it can up a bit to like 3,8gb...
- chummy
- Post #154
- Jan 18, 2016
- Forum: Graphics Cards
C
[Eurogamer] The state of 2GB VRAM GPUs

I made some tests because i want to understand better about that texture/vram usage. Ubisoft games works a bit different from GTA 5 in Vram management. I used custom low settings with the High texture quality(which require 2gb Vram in both games) to test in r9 390 and GTX560 1gb. When i...
- chummy
- Post #152
- Jan 18, 2016
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

Since in my situation the variations till nex gen is very obscure, i choose to get the 390. Here things are not so easy to say save 'X' during that time, because i do not earn in dollar or good currency, but devaluated one, which can devaluate more till next gen. I already was putting that in...
- chummy
- Post #26
- Nov 30, 2015
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

2 years, till end 2017 if i buy now.
- chummy
- Post #18
- Nov 30, 2015
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

The case i used to 2018 console was only based in a rumour, i know when new console generation arrives all things gonna change and we be forced to move on. So my words was not to suposedly 390 run next gen consoles(ps5), but keep working till this gen dies. I want a card to keep working at...
- chummy
- Post #14
- Nov 29, 2015
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

Sapphire Nitro OC. About $250 cards not coming better, how was possible to 960-380 at U$200 launch price reach 280x/770 which was U$330 at launch and both cards dont even get a litography change. Are you sure next gen $200~ cards will not get performance gains about 960-380 like before?
- chummy
- Post #6
- Nov 29, 2015
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

My other specs is 4770k@4.1ghz and Corsair RM650. I dont tell about the rest of system because i know they will keep fine with any single GPU. I play 1080p and i live in country which prices are not U$ but i make a comparison about what i probably can grab next gen. The 390 i have now to...
- chummy
- Post #3
- Nov 29, 2015
- Forum: Graphics Cards
C
Upgrade to R9 390 or wait 2016 next-gen..

Hello guys, i'm some reader in forum but start my first participation now. Let go to the discussion, the thing is curently i own a GTX770 2gb card and have option to purchase a r9 390. I know there upcoming new cards in somewhere 2016 but not exactly when. What i purchase next i want to...
- chummy
- Thread
- Nov 29, 2015
- Replies: 29
- Forum: Graphics Cards

RESOURCES

Top Bottom