COMPMID-1316 Using 8 bit dot product instruction in CLDepthWiseConvolution with QASYMM8
Change-Id: I3fc37bdceaae8b4b1effa51129b71bf352388564
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/138374
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
diff --git a/arm_compute/core/CL/CLHelpers.h b/arm_compute/core/CL/CLHelpers.h
index 3b025cc..ca1345d 100644
--- a/arm_compute/core/CL/CLHelpers.h
+++ b/arm_compute/core/CL/CLHelpers.h
@@ -110,6 +110,14 @@
*/
bool dot8_supported(const cl::Device &device);
+/** Helper function to check whether the cl_arm_integer_dot_product_accumulate_int8 extension is supported
+ *
+ * @param[in] device A CL device
+ *
+ * @return True if the extension is supported
+ */
+bool dot8_acc_supported(const cl::Device &device);
+
/** This function checks if the Winograd configuration (defined through the output tile, kernel size and the data layout) is supported on OpenCL
*
* @param[in] output_tile Output tile for the Winograd filtering algorithm