diff options
| -rw-r--r-- | cuda_base/README | 1 | ||||
| -rw-r--r-- | cuda_base/src/lib.rs | 17 | ||||
| -rw-r--r-- | zluda/README | 3 | ||||
| -rw-r--r-- | zluda/src/impl/context.rs | 25 | ||||
| -rw-r--r-- | zluda/src/impl/mod.rs | 20 | ||||
| -rw-r--r-- | zluda/src/lib.rs | 2 |
6 files changed, 37 insertions, 31 deletions
diff --git a/cuda_base/README b/cuda_base/README deleted file mode 100644 index 7ee6f45..0000000 --- a/cuda_base/README +++ /dev/null @@ -1 +0,0 @@ -bindgen build/wrapper.h -o src/cuda.rs --no-partialeq "CUDA_HOST_NODE_PARAMS_st" --with-derive-eq --allowlist-type="^CU.*" --allowlist-function="^cu.*" --allowlist-var="^CU.*" --default-enum-style=newtype --no-layout-tests --no-doc-comments --no-derive-debug --new-type-alias "^CUdevice_v\d+$|^CUdeviceptr_v\d+$" --must-use-type "cudaError_enum" --constified-enum "cudaError_enum" -- -I/usr/local/cuda/include
\ No newline at end of file diff --git a/cuda_base/src/lib.rs b/cuda_base/src/lib.rs index 366edd7..765af71 100644 --- a/cuda_base/src/lib.rs +++ b/cuda_base/src/lib.rs @@ -162,7 +162,13 @@ pub fn cuda_normalize_fn(tokens: TokenStream) -> TokenStream { .ident .to_string(); let known_modules = [ - "context", "device", "function", "link", "memory", "module", "pointer", + ("ctx", "context"), + ("device", "device"), + ("function", "function"), + ("link", "link"), + ("memory", "memory"), + ("module", "module"), + ("pointer", "pointer"), ]; let segments: Vec<String> = split(&fn_[2..]); let fn_path = join(segments, &known_modules); @@ -184,10 +190,13 @@ fn split(fn_: &str) -> Vec<String> { result } -fn join(fn_: Vec<String>, known_modules: &[&str]) -> Punctuated<Ident, Token![::]> { +fn join(fn_: Vec<String>, known_modules: &[(&str, &str)]) -> Punctuated<Ident, Token![::]> { let (prefix, suffix) = fn_.split_at(1); - if known_modules.contains(&&*prefix[0]) { - [&prefix[0], &suffix.join("_")] + if let Some((_, mod_name)) = known_modules + .iter() + .find(|(mod_prefix, _)| mod_prefix == &prefix[0]) + { + [*mod_name, &suffix.join("_")] .into_iter() .map(|seg| Ident::new(seg, Span::call_site())) .collect() diff --git a/zluda/README b/zluda/README deleted file mode 100644 index f6d929c..0000000 --- a/zluda/README +++ /dev/null @@ -1,3 +0,0 @@ -bindgen /usr/local/cuda/include/cuda.h -o cuda.rs --whitelist-function="^cu.*" --size_t-is-usize --default-enum-style=newtype --no-layout-tests --no-doc-comments --no-derive-debug --new-type-alias "^CUdevice$|^CUdeviceptr$" -sed -i -e 's/extern "C" {//g' -e 's/-> CUresult;/-> CUresult { impl_::unsupported()/g' -e 's/pub fn /#[no_mangle] pub extern "system" fn /g' cuda.rs -rustfmt cuda.rs
\ No newline at end of file diff --git a/zluda/src/impl/context.rs b/zluda/src/impl/context.rs index fffceb8..61cb92e 100644 --- a/zluda/src/impl/context.rs +++ b/zluda/src/impl/context.rs @@ -1,24 +1,9 @@ -use std::ptr; +use hip_runtime_sys::*; -use crate::cuda::CUlimit; -use crate::cuda::CUresult; - -pub(crate) unsafe fn get_limit(pvalue: *mut usize, limit: CUlimit) -> CUresult { - if pvalue == ptr::null_mut() { - return CUresult::CUDA_ERROR_INVALID_VALUE; - } - if limit == CUlimit::CU_LIMIT_STACK_SIZE { - *pvalue = 512; // GTX 1060 reports 1024 - CUresult::CUDA_SUCCESS - } else { - CUresult::CUDA_ERROR_NOT_SUPPORTED - } +pub(crate) unsafe fn get_limit(pvalue: *mut usize, limit: hipLimit_t) -> hipError_t { + unsafe { hipDeviceGetLimit(pvalue, limit) } } -pub(crate) fn set_limit(limit: CUlimit, value: usize) -> CUresult { - if limit == CUlimit::CU_LIMIT_STACK_SIZE { - CUresult::CUDA_SUCCESS - } else { - CUresult::CUDA_ERROR_NOT_SUPPORTED - } +pub(crate) fn set_limit(limit: hipLimit_t, value: usize) -> hipError_t { + unsafe { hipDeviceSetLimit(limit, value) } } diff --git a/zluda/src/impl/mod.rs b/zluda/src/impl/mod.rs index 8efd0a7..b2e8fc7 100644 --- a/zluda/src/impl/mod.rs +++ b/zluda/src/impl/mod.rs @@ -1,6 +1,7 @@ use cuda_types::*; use hip_runtime_sys::*; +pub(super) mod context; pub(super) mod device; #[cfg(debug_assertions)] @@ -17,7 +18,7 @@ pub(crate) trait FromCuda<'a, T>: Sized { fn from_cuda(t: &'a T) -> Result<Self, CUerror>; } -macro_rules! from_cuda_noop { +macro_rules! from_cuda_nop { ($($type_:ty),*) => { $( impl<'a> FromCuda<'a, $type_> for $type_ { @@ -65,18 +66,31 @@ macro_rules! from_cuda_transmute { }; } -from_cuda_noop!( +from_cuda_nop!( *mut i8, *mut usize, i32, u32, - cuda_types::CUdevprop, CUdevice_attribute + usize, + cuda_types::CUdevprop, + CUdevice_attribute ); from_cuda_transmute!( CUdevice => hipDevice_t, CUuuid => hipUUID ); +impl<'a> FromCuda<'a, CUlimit> for hipLimit_t { + fn from_cuda(limit: &'a CUlimit) -> Result<Self, CUerror> { + Ok(match *limit { + CUlimit::CU_LIMIT_STACK_SIZE => hipLimit_t::hipLimitStackSize, + CUlimit::CU_LIMIT_PRINTF_FIFO_SIZE => hipLimit_t::hipLimitPrintfFifoSize, + CUlimit::CU_LIMIT_MALLOC_HEAP_SIZE => hipLimit_t::hipLimitMallocHeapSize, + _ => return Err(CUerror::NOT_SUPPORTED), + }) + } +} + pub(crate) fn init(flags: ::core::ffi::c_uint) -> hipError_t { unsafe { hipInit(flags) } } diff --git a/zluda/src/lib.rs b/zluda/src/lib.rs index 7359e72..942c7e4 100644 --- a/zluda/src/lib.rs +++ b/zluda/src/lib.rs @@ -32,6 +32,8 @@ use cuda_base::cuda_function_declarations; cuda_function_declarations!( unimplemented, implemented <= [ + cuCtxGetLimit, + cuCtxSetLimit, cuDeviceComputeCapability, cuDeviceGet, cuDeviceGetAttribute, |