Ah so maybe it is a shader problem as i suspected.
Nice find, hardware.
Did you use Visual Studio to get that stack trace?
Ah so maybe it is a shader problem as i suspected.
The issue persists. Just reported in the GeForce drivers official forum.
Please guys, do the same and fill out the NVIDIA feedback form too, the more the better.
@Hardware Thank you! I hope NVIDIA listen to us … It would be nice if more affected users report to Nvidia too. As I said above, the more the better.
Agree that it can be a shader issue and I think it could have to do with the new graphic features/settings IOI added. Today and after several previous CTDs, I’ve been playing 6+ hours this afternoon without crashing with the new graphic settings all disabled.
Anyway, I’m going to do further test using different combinations of these new graphics settings to see what results I get.
Crashes happened before their new graphic settings. I like the shader theory.
I applied 60fps limit with RTSS, Gsync disabled, 1080p, crash again. I think we can not do anything on our side to stop that. But I found something interesting, just before the last crash, virtual memory jump from 12.5GB to ~18GB in 2 seconds, one second later game crashed. 18GB is the limit of my system (Virtual Memory = Physical memory (16GB) + Page file (2GB).
Crash again a few minutes ago, results :
10h31min59sec -> 13.3GB commited
10h32min01sec -> 17.1GB commited
On second later (10h32min02sec) => crash.
But I found something interesting, at first glance, the memory leak does not come directly from the game, the graph is stable :
@abeg Agreed, it seems unlinked with the new graphic settings after having tested all settings combos. Perhaps, what they do is that the error takes less time to occur. Shader bug theory still seems the most likely for me too… If you haven’t already done so, please report the issue on the official Nvidia forum and fill out the feedback form.
@Hardware Interesting results dude, how do you interpret them? Would you bet this is an error in how certain shaders interact/works with the Nvidia driver? If you can post on the Nvidia forum your latest findings and records.
I also have a RTX 2080 if that matters and i at least crash 2 or 3 times every hour… i recently bought the game and i cant belive after 1 month they still have such gamebreaking issue.
I do not know how DirectX pipeline and memory management works sorry. I am really tired of this issue. I removed my RTX2080 to put instead my old GTX970 back in my PC. Every day, I will update this post, until the next crash. But I can already say that my CPU is 20 to 30% less used than with my RTX2080 and the framerate is much more stable.
With GTX970 4GB :
- 12/23 : No crash
- 12/24 : No crash
- 12/25 : No crash
Mmm… I understand you the issue is weird and the situation disgusting but do you think the issue is at hardware level? … I’ve seen Pascal and some Maxwell users reporting the issue too.
my GTX 1080 didn’t have the issue, or maybe it’d have developed the issue with a game patch or drive update who knows… My nearly 3 year old 1080 packed in and left me thinking, I might as well use this as an excuse to upgrade
So I got myself a nice shiny new RTX 2080, and I was happy for a bit until I started playing Hitman 2, and thats when the crashing started.
It still crashes, I completely reinstalled Windows last night to have a nice fresh start, installed nothing else but the OS and the drivers and steam, played Hitman 2, crash to desktop with 0x80000003
Without Visual Studio installed it no longer takes 2-3 minutes to close.
The common factor is newer NVidia GPUs, which isn’t a lot of Hitman 2’s player base yet, but that will grow eventually, IO or NVidia will have to deal with this eventually.
Or you know what, just add in the DX12 backend, that’d probably solve it.
1080 owners also have this problem.
TacoFist has two 1080’s in SLI and he was seeing lots of 0x8000003 crashes.
With my 2080 and 417.35, the game is completely stable, except for:
- When I play Colorado - it crashes in a few minutes every time.
- When I use nVidia Ansel (any map).
I do not think it’s related to a specific generation or nvidia card model. It looks like the probability and crash frequence is proportional to the GPU performances, very few users report this problem with Maxwell cards or 1050/1050ti/1060, some with 1070/1080 and a lot with RTX 20x0 gen, so probably a software issue.
The game isn’t not stable and optimized atm with big GPU due to DX11 driver overhead, the number of draw call is limited by the CPU (classic case of bottleneck), which is explain why my CPU (i7 9700K) is much more used with my RTX2080 than with my GTX970.
With RTX2080, my performance are skewed towards the CPU. With my GTX970, the game skew the load back towards the GPU and increase performance output and stability.
Remember you want your GPU to be around 95%+ load, anything less you’ve got a CPU bottleneck, which is often the case with IOI’s Glacier 2 engine without DX12.
I wouldn’t use the adverb ‘completely’ in that case. The issue is still present.
It’s surely a Hitman 2 problem because i don’t have any issues with other games and 10 series cards and below are also having this issue…
I get the crash in any level, its just a matter of time.
I’ve had it in Miami, Most of the time in Chasing a Ghost and a few times in A New Life.
RTX 2080, latest drivers also.
Maybe ‘completely’ is the wrong word to use but I’ve tried hard to crash the game in the other maps on purpose to no avail. Only Colorado crashes for some reason.
Is it even possible to have a breakpoint in shader code?
For me, few times in Miami and Whittleton creek, most of the time in Mumbai and Santa Fortuna. Never on Sgail island and Hawke’s Bay.
I’ve had it happen in Hawkes Bay a few times. If it doesn’t crash i’ll say it again, you got lucky. Its a random bug, the worst kind of bug…