DigitalPrinceX
All performance issues because of AMD drivers, nice.

CruNcher
If that info is not official from developers, then ignore it, players don't know what is happening inside engine and you simply stuck somewhere.

Ok, another tests a bit modified 15 (compared to it), for amd cards, which better by stuttering and ctd (stil interesting in previous tests on nvidia cards).

EDIT: file deleted.

_________________
i5-4690k, 16Gb RAM, GTX 1060 6Gb, X-Fi Titanium, Win7 x128
I am INFP, not the brutal, godamnit.

Test 19 : Much better now. Similar to previous test 15.
Test 20 : Best performance of all tests today. Lesser stuttering. Average fps 25. Not the best but it is getting better.

When i get average fps 35 i call it success.
Before all fixes my average fps was 30 to 35

_________________
CPU :: Core i7 2600k @3.40GHz
GPU :: Sapphire Radeon HD 7870 GHz edition 2GB
Ram :: 8GB
Sound :: Creative X-Fi Titanium
OS :: Win 7 Ultimate 64-bit

For AMD cards now performance is the cost of reduced memory usage, even if i'll give you control under amount of memory freed, ctd will occur sooner, so i for performance and ctd both better to make other code. How about ctds with tests 19 vs 15? Which better (running longer and lower memory usage)?

_________________
i5-4690k, 16Gb RAM, GTX 1060 6Gb, X-Fi Titanium, Win7 x128
I am INFP, not the brutal, godamnit.

Test 19 runs longer. Test 20 runs more than 19.
Dont forget i am running more ugrid setting to push over the limit just to test. If i go default game will run more stable.

_________________
CPU :: Core i7 2600k @3.40GHz
GPU :: Sapphire Radeon HD 7870 GHz edition 2GB
Ram :: 8GB
Sound :: Creative X-Fi Titanium
OS :: Win 7 Ultimate 64-bit

If you are sure about 20 vs 19, then it's driver problem. Ok, better to make new variable in enblocal.ini for AMD, i can't force NVidia users to same hack. But what about 15 vs 19 or 20 by stability? They have some important differences.

_________________
i5-4690k, 16Gb RAM, GTX 1060 6Gb, X-Fi Titanium, Win7 x128
I am INFP, not the brutal, godamnit.

Пытался потестить но уже и не знаю как игру нагрузить, стоят HD текстуры, ugrids на 9, и производительность одинаковая что с 13713 что с тестовыми 193. ugrids 11 сам по себе очень нестабилен и вылетает где и как ему вздумается независимо от версий мода либо вовсе бесконечная загрузка при быстром путешествии\загрузке игры.
Попробую еще текстур понаустанавливать что-ли.

_________________
I used to be an ENB tweaker like you, but then I overheat my GPU.

Well i really like test 20. Went back to default ugrids and travel to places where i always CTD. No more crashes. I only crash if i console code spawn 20+ npc.
That is just me. Dont know how others are experiencing.

_________________
CPU :: Core i7 2600k @3.40GHz
GPU :: Sapphire Radeon HD 7870 GHz edition 2GB
Ram :: 8GB
Sound :: Creative X-Fi Titanium
OS :: Win 7 Ultimate 64-bit

DigitalPrinceX
Which from attached better than 20? These may have less stuttering, but more ctds.

EDIT: file deleted.

_________________
i5-4690k, 16Gb RAM, GTX 1060 6Gb, X-Fi Titanium, Win7 x128
I am INFP, not the brutal, godamnit.

Exact opposite for me on Nvidia.
Still using 314.22 drivers.
Even with default uGrids=5, CTD near Whiterun when running from Tundra.
X64 fix enabled or not, ReduceMemoryUsage enabled or not. (they're bypassed, right ?)

uGrids=7 makes immediate CTD when loading the game. (after savegame loading screen).

Test 19 or 20 the same.
No other memory tweak enabled in game inis than uGridsToLoad and uExteriorCellBuffer. (tried disabling this one as well, same results).
3GB or 4GB cards, the same.

_________________
Obsidian 900D, AX1500i, i7 6850K@4.4 Ghz, MSI X99 Raider, SLI 2x1080Ti 11G VRAM @2126 / 11720, 32G RAM@3361, 950 Pro M2, 11 Tb HDDs, W10 Pro 64, Full Watercooling


K ENB Pure Light

Oyama
Which is the best by stability and stuttering on NVidia?

EDIT: parameters forced inside, so don't touch them in enblocal.ini, also 15 vs 19/20?

_________________
i5-4690k, 16Gb RAM, GTX 1060 6Gb, X-Fi Titanium, Win7 x128
I am INFP, not the brutal, godamnit.