OutOfDeviceMemory on image.exe/teapot.exe GTX 770 4GB

[zpl-zak]

I get following error on my nvidia gtx 770-4gb when using examples. I'm using the latest nvidia drivers (368.69.0) on updated Windows 10 Home x64. Vulkan works on triangle.exe example.

W:\vulkano\examples\target\debug>teapot.exe
Using device: GTX 770 (type: DiscreteGpu)
thread '<main>' panicked at 'called `Result::unwrap()` on an `Err` value: OomError(OutOfDeviceMemory)', ../src/libcore\result.rs:746
note: Run with `RUST_BACKTRACE=1` for a backtrace.

W:\vulkano\examples\target\debug>triangle.exe
Using device: GTX 770 (type: DiscreteGpu)

W:\vulkano\examples\target\debug>image.exe
Using device: GTX 770 (type: DiscreteGpu)
thread '<main>' panicked at 'called `Result::unwrap()` on an `Err` value: OomError(OutOfDeviceMemory)', ../src/libcore\result.rs:746
note: Run with `RUST_BACKTRACE=1` for a backtrace.

[tomaka]

Could you run with RUST_BACKTRACE=1 (set RUST_BACKTRACE=1 from the command line) to see where the unwrap occurs?

[zpl-zak]

I've had to use gdb since Rust's backtrace doesn't show much:

(gdb) break rust_panic
Breakpoint 1 at 0x9c9140
(gdb) r
Starting program: W:\vulkano\examples\target\debug/teapot.exe
[New Thread 32956.0x8c74]
[New Thread 32956.0x8e4c]
[New Thread 32956.0x17b4]
[New Thread 32956.0x840c]
[New Thread 32956.0x8e04]
[New Thread 32956.0x8eb4]
[New Thread 32956.0x8eb0]
[New Thread 32956.0x8a44]
[New Thread 32956.0x6464]
Using device: GTX 770 (type: DiscreteGpu)
[New Thread 32956.0x8a58]
[New Thread 32956.0xbdc]
[New Thread 32956.0x884c]
[New Thread 32956.0x7260]
[New Thread 32956.0x8f04]
[New Thread 32956.0x8ed8]
[New Thread 32956.0x8ed0]
[New Thread 32956.0x8bd4]
[New Thread 32956.0x8f30]
[New Thread 32956.0x8138]
thread '<main>' panicked at 'called `Result::unwrap()` on an `Err` value: OomError(OutOfDeviceMemory)', ../src/libcore\result.rs:746
stack backtrace:
   0:   0x9e9c5d - <unknown>
   1:   0x9af2c1 - <unknown>
   2: 0x7756deca - <unknown>
   3:   0x9fd42a - <unknown>
   4:   0x9b0ec6 - <unknown>
   5:   0x435d61 - <unknown>
   6:   0x4021c3 - <unknown>
   7:   0x9e944d - <unknown>
   8:   0x50bab2 - <unknown>
   9:   0x4013dd - <unknown>
  10: 0x770d38f3 - <unknown>
  11: 0x77595de2 - <unknown>

Breakpoint 1, 0x009c9140 in rust_panic ()
(gdb) backtrace
#0  0x009c9140 in rust_panic ()
#1  0x009af467 in std::sys_common::unwind::begin_unwind_inner::h39d40f52add53ef7 ()
#2  0x009b0ec7 in std::sys_common::unwind::begin_unwind_fmt::h64c0ff793199cc1b ()
#3  0x009dd983 in rust_begin_unwind ()
#4  0x009fdf78 in core::panicking::panic_fmt::h73bf9d7e8e891a73 ()
#5  0x00438f26 in teapot::result::unwrap_failed<vulkano::image::sys::ImageCreationError> (msg=..., error=...)
    at ../src/libcore/result.rs:29
#6  0x00435d62 in teapot::result::{{impl}}::unwrap (self=...) at ../src/libcore/result.rs:687
#7  0x004021c4 in teapot::main () at src\bin/teapot.rs:69
#8  0x009e944e in std::sys_common::unwind::try::try_fn::h09ba69fd13531e58 ()
#9  0x009dd8fb in __rust_try ()
#10 0x009dd86c in std::sys_common::unwind::inner_try::h9eebd8dc83f388a6 ()
#11 0x009e92ba in std::rt::lang_start::h5b0863080165c75e ()
#12 0x0050bab3 in main () at ../src/libcore/mem.rs:114

Rust version:

1.9.0-i686-pc-windows-gnu
rustc 1.9.0 (e4e8b6668 2016-05-18)

[tomaka]

The stack trace leads to the moment where we create the transient depth buffer.
I suppose that's caused by lazily-allocated memory, but I'm not sure why.

[AustinJ235]

This seems related to #1247 closing in favor of that.