Leverage features of the mechanism to simplify implementation:
- The maximum number of possible cache-region reservations is a small
value known at compile-time
- Each reservation is unique (described by `CacheRegion`) so can have
a dedicated entry with either zero (free) or non-zero (reserved) value
So, there is no need for a dynamic collection (unordered_map here) to
keep track of reservations. A simple array is enough for that purpose.
Also, add some helper-code to enable array-indexing with the values of
`CacheRegion` enum.
Related-To: NEO-12837
Signed-off-by: Maciej Bielski <maciej.bielski@intel.com>
This PR handles the situation in which a component
has reserved a front window space for itself in the external heap,
so that the Compute Runtime cannot access this area.
In such a situation, we perform the following steps:
1. reserve 4GB chunk in heapStandard
2. split our chunk into 2 parts: heapFrontWindow, heapRegular
3. from this point on, map all linearStream allocations in reserved 4GB
chunk
Patch applies to Windows and WSL.
Patch only applies when the bindless global allocator is enabled.
Related-To: HSD-16025889919
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
Related-To: GSD-10253, GSD-9467, GSD-9381, NEO-11908
When EnableGlobalTimestampViaSubmission is set then
zeDeviceGetGlobalTimestamp uses immediate cmd submission
method to get GPU time.
Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
Add flag to log information for indirect detection debugging.
Add flag to disable indirect detection by kernel name.
Add flag to force indirect detection enable/disable for CM kernels.
Related-To: NEO-13372
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Performs minor renaming (mostly capitalization) in order to align with
specification.
Renames L1_CACHE_POLICY to L1_CACHE_CONTROL.
Related-To: NEO-13147
Signed-off-by: Vysochyn, Illia <illia.vysochyn@intel.com>
Change the error message for GPU page fault to match
with the message from gdb output
Related-To: NEO-13093
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
When using USM allocation, placement of BO can be simultaneously set to
local and system memory. This indicatates that migration between them
is allowed, but then we must not allow CPU caching to prevent corruptions.
Related-To: HSD-18040382956
Signed-off-by: Naklicki, Mateusz <mateusz.naklicki@intel.com>
Fix the PAT-index reporting in logger as currently on Windows reported
values are simply wrong.
The changed logic dependends on `RootDeviceEnvironment` and in order to
avoid introducing such dependencies into logger.[ch] the
`logAllocation()` is no longer a member of `FileLogger` but
a free-function instead (and a separate .cpp file). This is important
because the source files `logger.[ch]` are also used by ocloc library
and there is no point to contaminate ocloc code structure with
unnecessary dependencies.
Related-To: NEO-9421
Signed-off-by: Maciej Bielski <maciej.bielski@intel.com>
- don't load ze_loader.dll from file system
- to perform self-open on Windows use getModuleHandleA with proper module name
- don't free library loaded with getModuleHandleA
- as loader may be not available during runtime teardown:
- load translate handle function during global setup
- load setDriverTeardown function during global teardown
- when loader is not available during teardown, unset translate handle
function
Related-To: GSD-10147
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
on Windows use getModuleHandleA with proper module name
don't load ze_loader.dll from file system
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
