From 6b68096b04bd4323837a6fd29f68635b45c2f3d4 Mon Sep 17 00:00:00 2001
From: John Pennycook <john.pennycook@intel.com>
Date: Mon, 16 Dec 2024 10:18:26 +0000
Subject: [PATCH] [SYCL][Doc] Add slm_per_subslice query for Xe GPUs

Extends sycl_ext_intel_device_info with a new query for the amount of SLM per
subslice. The amount of SLM per subslice may be greater than the amount of
memory available to a single work-group, and can be used to reason about
occupancy.

Signed-off-by: John Pennycook <john.pennycook@intel.com>
---
 .../supported/sycl_ext_intel_device_info.md   | 43 +++++++++++++++++++
 1 file changed, 43 insertions(+)

diff --git a/sycl/doc/extensions/supported/sycl_ext_intel_device_info.md b/sycl/doc/extensions/supported/sycl_ext_intel_device_info.md
index 7e62a781cdf0d..62a89eac0512c 100644
--- a/sycl/doc/extensions/supported/sycl_ext_intel_device_info.md
+++ b/sycl/doc/extensions/supported/sycl_ext_intel_device_info.md
@@ -19,6 +19,7 @@ The Feature Test Macro SYCL\_EXT\_INTEL\_DEVICE\_INFO will be defined as one of
 | 4     | Free device memory query is supported |
 | 5     | Device ID is supported |
 | 6     | Memory clock rate and bus width queries are supported |
+| 7     | SLM per subslice is supported |
 
 
 
@@ -505,6 +506,48 @@ Then the memory bus width can be obtained using the standard get\_info() interfa
       auto MemoryBusWidth = dev.get_info<ext::intel::info::device::memory_bus_width>();
     }
 
+# Intel GPU SLM per Subslice #
+
+A new device descriptor will be added which will provide the amount of shared
+local memory (SLM) per subslice on an Intel GPU.
+
+**Note:** The amount of SLM per subslice may be greater than the amount of
+memory available to a single work-group, and can be used to reason about
+occupancy.
+
+
+## Version ##
+
+The extension supports this query in version 7 and later.
+
+
+## Device Information Descriptors ##
+
+| Device Descriptors | Return Type | Description |
+| ------------------ | ----------- | ----------- |
+| ext\:\:intel\:\:info\:\:device\:\:gpu\_slm\_per\_subslice | uint64\_t| Returns the amount of SLM per subslice, in bytes. |
+
+
+## Aspects ##
+
+A new aspect, ext\_intel\_gpu\_slm\_per\_subslice, will be added.
+
+
+## Error Condition ##
+
+Throws a synchronous `exception` with the `errc::feature_not_supported` error
+code if the device does not have `aspect::ext_intel_gpu_slm_per_subslice`.
+
+## Example Usage ##
+
+The amount of SLM per subslice can be obtained using the standard get\_info()
+interface.
+
+    if (dev.has(aspect::ext_intel_gpu_slm_per_subslice)) {
+      auto slmBytes = dev.get_info<ext::intel::info::device::gpu_slm_per_subslice>();
+    }
+
+
 # Deprecated queries #
 
 The table below lists deprecated, that would soon be removed and their replacements: