static const struct brw_device_info brw_device_info_i965 = {
.gen = 4,
.has_negative_rhw_bug = true,
+ .num_slices = 1,
.max_vs_threads = 16,
.max_gs_threads = 2,
.max_wm_threads = 8 * 4,
.has_compr4 = true,
.has_surface_tile_offset = true,
.is_g4x = true,
+ .num_slices = 1,
.max_vs_threads = 32,
.max_gs_threads = 2,
.max_wm_threads = 10 * 5,
.has_pln = true,
.has_compr4 = true,
.has_surface_tile_offset = true,
+ .num_slices = 1,
.max_vs_threads = 72,
.max_gs_threads = 32,
.max_wm_threads = 12 * 6,
.has_pln = true,
.has_surface_tile_offset = true,
.needs_unlit_centroid_workaround = true,
+ .num_slices = 1,
.max_vs_threads = 24,
.max_gs_threads = 21, /* conservative; 24 if rendering disabled. */
.max_wm_threads = 40,
.has_pln = true,
.has_surface_tile_offset = true,
.needs_unlit_centroid_workaround = true,
+ .num_slices = 1,
.max_vs_threads = 60,
.max_gs_threads = 60,
.max_wm_threads = 80,
static const struct brw_device_info brw_device_info_ivb_gt1 = {
GEN7_FEATURES, .is_ivybridge = true, .gt = 1,
.needs_unlit_centroid_workaround = true,
+ .num_slices = 1,
.max_vs_threads = 36,
.max_hs_threads = 36,
.max_ds_threads = 36,
.max_gs_threads = 36,
.max_wm_threads = 48,
+ .max_cs_threads = 36,
.urb = {
.size = 128,
.min_vs_entries = 32,
.max_vs_entries = 512,
.max_hs_entries = 32,
+ .min_ds_entries = 10,
.max_ds_entries = 288,
.max_gs_entries = 192,
},
static const struct brw_device_info brw_device_info_ivb_gt2 = {
GEN7_FEATURES, .is_ivybridge = true, .gt = 2,
.needs_unlit_centroid_workaround = true,
+ .num_slices = 1,
.max_vs_threads = 128,
.max_hs_threads = 128,
.max_ds_threads = 128,
.max_gs_threads = 128,
.max_wm_threads = 172,
+ .max_cs_threads = 64,
.urb = {
.size = 256,
.min_vs_entries = 32,
.max_vs_entries = 704,
.max_hs_entries = 64,
+ .min_ds_entries = 10,
.max_ds_entries = 448,
.max_gs_entries = 320,
},
static const struct brw_device_info brw_device_info_byt = {
GEN7_FEATURES, .is_baytrail = true, .gt = 1,
.needs_unlit_centroid_workaround = true,
+ .num_slices = 1,
.has_llc = false,
.max_vs_threads = 36,
.max_hs_threads = 36,
.max_ds_threads = 36,
.max_gs_threads = 36,
.max_wm_threads = 48,
+ .max_cs_threads = 32,
.urb = {
.size = 128,
.min_vs_entries = 32,
.max_vs_entries = 512,
.max_hs_entries = 32,
+ .min_ds_entries = 10,
.max_ds_entries = 288,
.max_gs_entries = 192,
},
};
+#define HSW_FEATURES \
+ GEN7_FEATURES, \
+ .is_haswell = true, \
+ .supports_simd16_3src = true, \
+ .has_resource_streamer = true
+
static const struct brw_device_info brw_device_info_hsw_gt1 = {
- GEN7_FEATURES, .is_haswell = true, .gt = 1,
+ HSW_FEATURES, .gt = 1,
+ .num_slices = 1,
.max_vs_threads = 70,
.max_hs_threads = 70,
.max_ds_threads = 70,
.max_gs_threads = 70,
.max_wm_threads = 102,
+ .max_cs_threads = 70,
.urb = {
.size = 128,
.min_vs_entries = 32,
.max_vs_entries = 640,
.max_hs_entries = 64,
+ .min_ds_entries = 10,
.max_ds_entries = 384,
.max_gs_entries = 256,
},
};
static const struct brw_device_info brw_device_info_hsw_gt2 = {
- GEN7_FEATURES, .is_haswell = true, .gt = 2,
+ HSW_FEATURES, .gt = 2,
+ .num_slices = 1,
.max_vs_threads = 280,
.max_hs_threads = 256,
.max_ds_threads = 280,
.max_gs_threads = 256,
.max_wm_threads = 204,
+ .max_cs_threads = 70,
.urb = {
.size = 256,
.min_vs_entries = 64,
.max_vs_entries = 1664,
.max_hs_entries = 128,
+ .min_ds_entries = 10,
.max_ds_entries = 960,
.max_gs_entries = 640,
},
};
static const struct brw_device_info brw_device_info_hsw_gt3 = {
- GEN7_FEATURES, .is_haswell = true, .gt = 3,
+ HSW_FEATURES, .gt = 3,
+ .num_slices = 2,
.max_vs_threads = 280,
.max_hs_threads = 256,
.max_ds_threads = 280,
.max_gs_threads = 256,
.max_wm_threads = 408,
+ .max_cs_threads = 70,
.urb = {
.size = 512,
.min_vs_entries = 64,
.max_vs_entries = 1664,
.max_hs_entries = 128,
+ .min_ds_entries = 10,
.max_ds_entries = 960,
.max_gs_entries = 640,
},
#define GEN8_FEATURES \
.gen = 8, \
.has_hiz_and_separate_stencil = true, \
+ .has_resource_streamer = true, \
.must_use_separate_stencil = true, \
.has_llc = true, \
.has_pln = true, \
+ .supports_simd16_3src = true, \
.max_vs_threads = 504, \
.max_hs_threads = 504, \
.max_ds_threads = 504, \
.max_gs_threads = 504, \
- .max_wm_threads = 384 \
+ .max_wm_threads = 384
static const struct brw_device_info brw_device_info_bdw_gt1 = {
GEN8_FEATURES, .gt = 1,
+ .num_slices = 1,
+ .max_cs_threads = 42,
.urb = {
.size = 192,
.min_vs_entries = 64,
.max_vs_entries = 2560,
.max_hs_entries = 504,
+ .min_ds_entries = 34,
.max_ds_entries = 1536,
.max_gs_entries = 960,
}
static const struct brw_device_info brw_device_info_bdw_gt2 = {
GEN8_FEATURES, .gt = 2,
+ .num_slices = 1,
+ .max_cs_threads = 56,
.urb = {
.size = 384,
.min_vs_entries = 64,
.max_vs_entries = 2560,
.max_hs_entries = 504,
+ .min_ds_entries = 34,
.max_ds_entries = 1536,
.max_gs_entries = 960,
}
static const struct brw_device_info brw_device_info_bdw_gt3 = {
GEN8_FEATURES, .gt = 3,
+ .num_slices = 2,
+ .max_cs_threads = 56,
.urb = {
.size = 384,
.min_vs_entries = 64,
.max_vs_entries = 2560,
.max_hs_entries = 504,
+ .min_ds_entries = 34,
.max_ds_entries = 1536,
.max_gs_entries = 960,
}
};
-/* Thread counts and URB limits are placeholders, and may not be accurate.
- * These were copied from Haswell GT1, above.
- */
static const struct brw_device_info brw_device_info_chv = {
GEN8_FEATURES, .is_cherryview = 1, .gt = 1,
.has_llc = false,
+ .num_slices = 1,
.max_vs_threads = 80,
.max_hs_threads = 80,
.max_ds_threads = 80,
.max_gs_threads = 80,
.max_wm_threads = 128,
+ .max_cs_threads = 28,
.urb = {
- .size = 128,
+ .size = 192,
.min_vs_entries = 34,
.max_vs_entries = 640,
.max_hs_entries = 80,
+ .min_ds_entries = 34,
.max_ds_entries = 384,
.max_gs_entries = 256,
}
};
-/* Thread counts and URB limits are placeholders, and may not be accurate. */
#define GEN9_FEATURES \
.gen = 9, \
.has_hiz_and_separate_stencil = true, \
+ .has_resource_streamer = true, \
.must_use_separate_stencil = true, \
.has_llc = true, \
.has_pln = true, \
- .max_vs_threads = 280, \
- .max_gs_threads = 256, \
- .max_wm_threads = 408, \
+ .supports_simd16_3src = true, \
+ .max_vs_threads = 336, \
+ .max_gs_threads = 336, \
+ .max_hs_threads = 336, \
+ .max_ds_threads = 336, \
+ .max_wm_threads = 64 * 9, \
+ .max_cs_threads = 56, \
.urb = { \
- .size = 128, \
+ .size = 384, \
.min_vs_entries = 64, \
- .max_vs_entries = 1664, \
+ .max_vs_entries = 1856, \
+ .max_hs_entries = 672, \
+ .min_ds_entries = 34, \
+ .max_ds_entries = 1120, \
.max_gs_entries = 640, \
}
static const struct brw_device_info brw_device_info_skl_gt1 = {
- GEN9_FEATURES, .gt = 1
+ GEN9_FEATURES, .gt = 1,
+ .num_slices = 1,
+ .urb.size = 192,
};
static const struct brw_device_info brw_device_info_skl_gt2 = {
- GEN9_FEATURES, .gt = 2
+ GEN9_FEATURES, .gt = 2,
+ .num_slices = 1,
};
static const struct brw_device_info brw_device_info_skl_gt3 = {
- GEN9_FEATURES, .gt = 3
+ GEN9_FEATURES, .gt = 3,
+ .num_slices = 2,
+};
+
+static const struct brw_device_info brw_device_info_skl_gt4 = {
+ GEN9_FEATURES, .gt = 4,
+ .num_slices = 3,
+ /* From the "L3 Allocation and Programming" documentation:
+ *
+ * "URB is limited to 1008KB due to programming restrictions. This is not a
+ * restriction of the L3 implementation, but of the FF and other clients.
+ * Therefore, in a GT4 implementation it is possible for the programmed
+ * allocation of the L3 data array to provide 3*384KB=1152KB for URB, but
+ * only 1008KB of this will be used."
+ */
+ .urb.size = 1008 / 3,
+};
+
+static const struct brw_device_info brw_device_info_bxt = {
+ GEN9_FEATURES,
+ .is_broxton = 1,
+ .gt = 1,
+ .has_llc = false,
+
+ /* XXX: These are preliminary thread counts and URB sizes. */
+ .num_slices = 1,
+ .max_vs_threads = 56,
+ .max_hs_threads = 56,
+ .max_ds_threads = 56,
+ .max_gs_threads = 56,
+ .max_wm_threads = 32,
+ .max_cs_threads = 28,
+ .urb = {
+ .size = 64,
+ .min_vs_entries = 34,
+ .max_vs_entries = 640,
+ .max_hs_entries = 80,
+ .max_ds_entries = 80,
+ .max_gs_entries = 256,
+ }
+};
+
+/*
+ * Note: for all KBL SKUs, the PRM says SKL for GS entries, not SKL+.
+ * There's no KBL entry. Using the default SKL (GEN9) GS entries value.
+ */
+
+/*
+ * Both SKL and KBL support a maximum of 64 threads per
+ * Pixel Shader Dispatch (PSD) unit.
+ */
+#define KBL_MAX_THREADS_PER_PSD 64
+
+static const struct brw_device_info brw_device_info_kbl_gt1 = {
+ GEN9_FEATURES,
+ .gt = 1,
+
+ .max_cs_threads = 7 * 6,
+ .max_wm_threads = KBL_MAX_THREADS_PER_PSD * 2,
+ .urb.size = 192,
+ .num_slices = 1,
+};
+
+static const struct brw_device_info brw_device_info_kbl_gt1_5 = {
+ GEN9_FEATURES,
+ .gt = 1,
+
+ .max_cs_threads = 7 * 6,
+ .max_wm_threads = KBL_MAX_THREADS_PER_PSD * 3,
+ .num_slices = 1,
+};
+
+static const struct brw_device_info brw_device_info_kbl_gt2 = {
+ GEN9_FEATURES,
+ .gt = 2,
+
+ .max_wm_threads = KBL_MAX_THREADS_PER_PSD * 3,
+ .num_slices = 1,
+};
+
+static const struct brw_device_info brw_device_info_kbl_gt3 = {
+ GEN9_FEATURES,
+ .gt = 3,
+
+ .max_wm_threads = KBL_MAX_THREADS_PER_PSD * 6,
+ .num_slices = 2,
+};
+
+static const struct brw_device_info brw_device_info_kbl_gt4 = {
+ GEN9_FEATURES,
+ .gt = 4,
+
+ .max_wm_threads = KBL_MAX_THREADS_PER_PSD * 9,
+ /*
+ * From the "L3 Allocation and Programming" documentation:
+ *
+ * "URB is limited to 1008KB due to programming restrictions. This
+ * is not a restriction of the L3 implementation, but of the FF and
+ * other clients. Therefore, in a GT4 implementation it is
+ * possible for the programmed allocation of the L3 data array to
+ * provide 3*384KB=1152KB for URB, but only 1008KB of this
+ * will be used."
+ */
+ .urb.size = 1008 / 3,
+ .num_slices = 3,
};
const struct brw_device_info *
brw_get_device_info(int devid)
{
+ const struct brw_device_info *devinfo;
switch (devid) {
#undef CHIPSET
-#define CHIPSET(id, family, name) case id: return &brw_device_info_##family;
+#define CHIPSET(id, family, name) \
+ case id: devinfo = &brw_device_info_##family; break;
#include "pci_ids/i965_pci_ids.h"
default:
fprintf(stderr, "i965_dri.so does not support the 0x%x PCI ID.\n", devid);
return NULL;
}
+
+ return devinfo;
}