From: Matthew Brost <matthew.brost@intel.com>
To: intel-xe@lists.freedesktop.org
Cc: stuart.summers@intel.com, arvind.yadav@intel.com,
himal.prasad.ghimiray@intel.com,
thomas.hellstrom@linux.intel.com, francois.dugast@intel.com
Subject: [PATCH v3 00/25] CPU binds and ULLS on migration queue
Date: Fri, 27 Feb 2026 17:34:36 -0800 [thread overview]
Message-ID: <20260228013501.106680-1-matthew.brost@intel.com> (raw)
We now have data demonstrating the need for CPU binds and ULLS on the
migration queue, based on results generated from [1].
On BMG, measurements show that when the GPU is continuously processing
faults, copy jobs run approximately 30–40µs faster (depending on the
test case) with ULLS compared to traditional GuC submission with SLPC
enabled on the migration queue. Startup from a cold GPU shows an even
larger speedup. Given the critical nature of fault performance, ULLS
appears to be a worthwhile feature.
In addition to driver telemetry, UMD compute benchmarks consistently
show over 1GB/s improvement in pagefault benchmarks with ULLS enabled.
ULLS will consume more power (not yet measured) due to a continuously
running batch on the paging engine. However, compute UMDs already do
this on engines exposed to users, so this seems like a worthwhile
tradeoff. To mitigate power concerns, ULLS will exit after a period of
time in which no faults have been processed.
CPU binds are required for ULLS to function, as the migration queue
needs exclusive access to the paging hardware engine. Thus, CPU binds
are included here.
Beyond being a requirement for ULLS, CPU binds should also reduce
VM-bind latency, provide clearer multi-tile and TLB-invalidation
layering, reduce pressure on GuC during fault storms as it is bypassed,
and decouple kernel binds from unrelated copy/clear jobs—especially
beneficial when faults are serviced in parallel. In a parallel-faulting
test case, average bind time was reduced by approximately 15µs. In the
worst case, 2MB copy time (~60–140µs) × (number of pagefault threads −
1) of latency would otherwise be added to a single fault. Reducing this
latency increases overall throughput of the fault handler.
This series can be merged in phases:
Phase 1: CPU binds (patches 1–13)
Phase 2: CPU-bind components and multi-tile relayers (patches 14–17)
Phase 3: ULLS on the migration execution queue (patches 18–25)
v2:
- Use delayed worker to exit ULLS mode in an effort to save on power
- Various other cleanups
v3:
- CPU bind component, multi-tile relayer
- Split CPU bind patches in many small patches
Matt
[1] https://patchwork.freedesktop.org/series/149811/
Matthew Brost (25):
drm/xe: Drop struct xe_migrate_pt_update argument from populate/clear
vfuns
drm/xe: Add xe_migrate_update_pgtables_cpu_execute helper
drm/xe: Decouple exec queue idle check from LRC
drm/xe: Add job count to GuC exec queue snapshot
drm/xe: Update xe_bo_put_deferred arguments to include writeback flag
drm/xe: Add XE_BO_FLAG_PUT_VM_ASYNC
drm/xe: Update scheduler job layer to support PT jobs
drm/xe: Add helpers to access PT ops
drm/xe: Add struct xe_pt_job_ops
drm/xe: Update GuC submission backend to run PT jobs
drm/xe: Store level in struct xe_vm_pgtable_update
drm/xe: Don't use migrate exec queue for page fault binds
drm/xe: Enable CPU binds for jobs
drm/xe: Remove unused arguments from xe_migrate_pt_update_ops
drm/xe: Make bind queues operate cross-tile
drm/xe: Add CPU bind layer
drm/xe: Add device flag to enable PT mirroring across tiles
drm/xe: Add xe_hw_engine_write_ring_tail
drm/xe: Add ULLS support to LRC
drm/xe: Add ULLS migration job support to migration layer
drm/xe: Add MI_SEMAPHORE_WAIT instruction defs
drm/xe: Add ULLS migration job support to ring ops
drm/xe: Add ULLS migration job support to GuC submission
drm/xe: Enter ULLS for migration jobs upon page fault or SVM prefetch
drm/xe: Add modparam to enable / disable ULLS on migrate queue
drivers/gpu/drm/xe/Makefile | 1 +
.../gpu/drm/xe/instructions/xe_mi_commands.h | 6 +
drivers/gpu/drm/xe/xe_bo.c | 8 +-
drivers/gpu/drm/xe/xe_bo.h | 11 +-
drivers/gpu/drm/xe/xe_bo_types.h | 2 -
drivers/gpu/drm/xe/xe_cpu_bind.c | 296 +++++++
drivers/gpu/drm/xe/xe_cpu_bind.h | 118 +++
drivers/gpu/drm/xe/xe_debugfs.c | 1 +
drivers/gpu/drm/xe/xe_defaults.h | 1 +
drivers/gpu/drm/xe/xe_device.c | 17 +-
drivers/gpu/drm/xe/xe_device_types.h | 11 +
drivers/gpu/drm/xe/xe_drm_client.c | 2 +-
drivers/gpu/drm/xe/xe_exec_queue.c | 163 ++--
drivers/gpu/drm/xe/xe_exec_queue.h | 18 +-
drivers/gpu/drm/xe/xe_exec_queue_types.h | 21 +-
drivers/gpu/drm/xe/xe_guc_submit.c | 82 +-
drivers/gpu/drm/xe/xe_guc_submit_types.h | 2 +
drivers/gpu/drm/xe/xe_hw_engine.c | 10 +
drivers/gpu/drm/xe/xe_hw_engine.h | 1 +
drivers/gpu/drm/xe/xe_lrc.c | 51 ++
drivers/gpu/drm/xe/xe_lrc.h | 3 +
drivers/gpu/drm/xe/xe_lrc_types.h | 4 +
drivers/gpu/drm/xe/xe_migrate.c | 585 +++++--------
drivers/gpu/drm/xe/xe_migrate.h | 93 +--
drivers/gpu/drm/xe/xe_module.c | 4 +
drivers/gpu/drm/xe/xe_module.h | 1 +
drivers/gpu/drm/xe/xe_pagefault.c | 3 +
drivers/gpu/drm/xe/xe_pci.c | 2 +
drivers/gpu/drm/xe/xe_pci_types.h | 1 +
drivers/gpu/drm/xe/xe_pt.c | 773 +++++++++++-------
drivers/gpu/drm/xe/xe_pt.h | 12 +-
drivers/gpu/drm/xe/xe_pt_types.h | 49 +-
drivers/gpu/drm/xe/xe_ring_ops.c | 31 +
drivers/gpu/drm/xe/xe_sched_job.c | 100 ++-
drivers/gpu/drm/xe/xe_sched_job_types.h | 36 +-
drivers/gpu/drm/xe/xe_sync.c | 20 +-
drivers/gpu/drm/xe/xe_tlb_inval_job.c | 28 +-
drivers/gpu/drm/xe/xe_tlb_inval_job.h | 4 +-
drivers/gpu/drm/xe/xe_vm.c | 241 +++---
drivers/gpu/drm/xe/xe_vm.h | 3 +
drivers/gpu/drm/xe/xe_vm_types.h | 22 +-
41 files changed, 1658 insertions(+), 1179 deletions(-)
create mode 100644 drivers/gpu/drm/xe/xe_cpu_bind.c
create mode 100644 drivers/gpu/drm/xe/xe_cpu_bind.h
--
2.34.1
next reply other threads:[~2026-02-28 1:35 UTC|newest]
Thread overview: 63+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-02-28 1:34 Matthew Brost [this message]
2026-02-28 1:34 ` [PATCH v3 01/25] drm/xe: Drop struct xe_migrate_pt_update argument from populate/clear vfuns Matthew Brost
2026-03-05 14:17 ` Francois Dugast
2026-02-28 1:34 ` [PATCH v3 02/25] drm/xe: Add xe_migrate_update_pgtables_cpu_execute helper Matthew Brost
2026-03-05 14:39 ` Francois Dugast
2026-02-28 1:34 ` [PATCH v3 03/25] drm/xe: Decouple exec queue idle check from LRC Matthew Brost
2026-03-02 20:50 ` Summers, Stuart
2026-03-02 21:02 ` Matthew Brost
2026-03-03 21:26 ` Summers, Stuart
2026-03-03 22:42 ` Matthew Brost
2026-03-03 22:54 ` Summers, Stuart
2026-02-28 1:34 ` [PATCH v3 04/25] drm/xe: Add job count to GuC exec queue snapshot Matthew Brost
2026-03-02 20:50 ` Summers, Stuart
2026-02-28 1:34 ` [PATCH v3 05/25] drm/xe: Update xe_bo_put_deferred arguments to include writeback flag Matthew Brost
2026-04-01 12:20 ` Francois Dugast
2026-04-01 22:39 ` Matthew Brost
2026-02-28 1:34 ` [PATCH v3 06/25] drm/xe: Add XE_BO_FLAG_PUT_VM_ASYNC Matthew Brost
2026-04-01 12:22 ` Francois Dugast
2026-04-01 22:38 ` Matthew Brost
2026-02-28 1:34 ` [PATCH v3 07/25] drm/xe: Update scheduler job layer to support PT jobs Matthew Brost
2026-03-03 22:50 ` Summers, Stuart
2026-03-03 23:00 ` Matthew Brost
2026-02-28 1:34 ` [PATCH v3 08/25] drm/xe: Add helpers to access PT ops Matthew Brost
2026-04-07 15:22 ` Francois Dugast
2026-02-28 1:34 ` [PATCH v3 09/25] drm/xe: Add struct xe_pt_job_ops Matthew Brost
2026-03-03 23:26 ` Summers, Stuart
2026-03-03 23:28 ` Matthew Brost
2026-02-28 1:34 ` [PATCH v3 10/25] drm/xe: Update GuC submission backend to run PT jobs Matthew Brost
2026-03-03 23:28 ` Summers, Stuart
2026-03-04 0:26 ` Matthew Brost
2026-03-04 20:43 ` Summers, Stuart
2026-03-04 21:53 ` Matthew Brost
2026-03-05 20:24 ` Summers, Stuart
2026-02-28 1:34 ` [PATCH v3 11/25] drm/xe: Store level in struct xe_vm_pgtable_update Matthew Brost
2026-03-03 23:44 ` Summers, Stuart
2026-02-28 1:34 ` [PATCH v3 12/25] drm/xe: Don't use migrate exec queue for page fault binds Matthew Brost
2026-02-28 1:34 ` [PATCH v3 13/25] drm/xe: Enable CPU binds for jobs Matthew Brost
2026-02-28 1:34 ` [PATCH v3 14/25] drm/xe: Remove unused arguments from xe_migrate_pt_update_ops Matthew Brost
2026-02-28 1:34 ` [PATCH v3 15/25] drm/xe: Make bind queues operate cross-tile Matthew Brost
2026-02-28 1:34 ` [PATCH v3 16/25] drm/xe: Add CPU bind layer Matthew Brost
2026-02-28 1:34 ` [PATCH v3 17/25] drm/xe: Add device flag to enable PT mirroring across tiles Matthew Brost
2026-02-28 1:34 ` [PATCH v3 18/25] drm/xe: Add xe_hw_engine_write_ring_tail Matthew Brost
2026-02-28 1:34 ` [PATCH v3 19/25] drm/xe: Add ULLS support to LRC Matthew Brost
2026-03-05 20:21 ` Francois Dugast
2026-02-28 1:34 ` [PATCH v3 20/25] drm/xe: Add ULLS migration job support to migration layer Matthew Brost
2026-03-05 23:34 ` Summers, Stuart
2026-03-09 23:11 ` Matthew Brost
2026-02-28 1:34 ` [PATCH v3 21/25] drm/xe: Add MI_SEMAPHORE_WAIT instruction defs Matthew Brost
2026-02-28 1:34 ` [PATCH v3 22/25] drm/xe: Add ULLS migration job support to ring ops Matthew Brost
2026-02-28 1:34 ` [PATCH v3 23/25] drm/xe: Add ULLS migration job support to GuC submission Matthew Brost
2026-02-28 1:35 ` [PATCH v3 24/25] drm/xe: Enter ULLS for migration jobs upon page fault or SVM prefetch Matthew Brost
2026-02-28 1:35 ` [PATCH v3 25/25] drm/xe: Add modparam to enable / disable ULLS on migrate queue Matthew Brost
2026-03-05 22:59 ` Summers, Stuart
2026-04-01 22:44 ` Matthew Brost
2026-02-28 1:43 ` ✗ CI.checkpatch: warning for CPU binds and ULLS on migration queue (rev3) Patchwork
2026-02-28 1:44 ` ✓ CI.KUnit: success " Patchwork
2026-02-28 2:32 ` ✓ Xe.CI.BAT: " Patchwork
2026-02-28 13:59 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-03-02 17:54 ` Summers, Stuart
2026-03-02 18:13 ` Matthew Brost
2026-03-05 22:56 ` [PATCH v3 00/25] CPU binds and ULLS on migration queue Summers, Stuart
2026-03-10 22:17 ` Matthew Brost
2026-03-20 15:31 ` Thomas Hellström
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260228013501.106680-1-matthew.brost@intel.com \
--to=matthew.brost@intel.com \
--cc=arvind.yadav@intel.com \
--cc=francois.dugast@intel.com \
--cc=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=stuart.summers@intel.com \
--cc=thomas.hellstrom@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox