RE: [PATCH v5] enhance NUMA affinity heuristic

> -Original Message- > From: Thomas Monjalon > Sent: 2023年4月21日 16:13 > To: You, KaisenX > Cc: dev@dpdk.org; Zhou, YidingX ; > david.march...@redhat.com; Matz, Olivier ; > ferruh.yi...@amd.com; zhou...@loongson.cn; sta...@dpdk.org; > Richardson, Bruce ; jer...@marvell.com; > Burakov, A

[PATCH v1 5/5] ml/cnxk: implement xstats driver functions

Added support for revised xstats APIs in cnxk ML driver. Signed-off-by: Srikanth Yalavarthi --- doc/guides/mldevs/cnxk.rst | 30 +- drivers/ml/cnxk/cn10k_ml_dev.h | 96 +- drivers/ml/cnxk/cn10k_ml_model.h | 21 -- drivers/ml/cnxk/cn10k_ml_ops.c | 530 ++--

[PATCH v1 4/5] app/mldev: enable reporting xstats

Enabled reporting xstats in ML test application. Enabled stats option for model_ops test case. Added common files for xstats and throughput functions. Signed-off-by: Srikanth Yalavarthi --- app/test-mldev/meson.build | 1 + app/test-mldev/ml_common.h | 11 ++ a

[PATCH v1 3/5] mldev: implement xstats library functions

Implemented xstats library functions as per revised spec. Signed-off-by: Srikanth Yalavarthi --- lib/mldev/rte_mldev.c | 91 +++ lib/mldev/rte_mldev_core.h | 107 + lib/mldev/version.map | 4 ++ 3 files changed, 202 ins

[PATCH v1 2/5] mldev: introduce revised xstats

Introduce revised xstats APIs to support reporting device and per-model xstats. Stat type is selected through mode parameter. Support modes include device and model. Signed-off-by: Srikanth Yalavarthi --- lib/mldev/rte_mldev.h | 113 ++ 1 file changed, 113

[PATCH v1 1/5] mldev: remove xstats APIs from library

This change is a preparatoy step for revised xstats APIs. Revised xstats APIs support reporting device and per model stats, which is based on eventdev xstats. Removed xstats APIs from spec and library implementation. Disabled reporting xstats in test application and disabled xstats functions in dr

[PATCH v1 0/5] Implementation of revised ML xstats spec

This series of patches introduces revised xstats specification for ML device. The revised xstats spec is based on eventdev xstats and supports DEVICE and MODEL modes to get xstats. This enables retrieving xstats for device and each model separately. Srikanth Yalavarthi (5): mldev: remove xstats

[PATCH v1 3/3] ml/cnxk: add support for 32 I/O per model

Added support for 32 inputs and outputs per model. Signed-off-by: Srikanth Yalavarthi --- drivers/ml/cnxk/cn10k_ml_model.c | 374 ++- drivers/ml/cnxk/cn10k_ml_model.h | 5 +- drivers/ml/cnxk/cn10k_ml_ops.c | 125 --- 3 files changed, 367 insertions(+), 137

[PATCH v1 2/3] ml/cnxk: update model metadata to v2301

Update model metadata to v2301. Revised metadata introduces fields to support up to 32 inputs/outputs per model, scratch relocation and updates to names of existing fields. Update driver files to include changes in names of metadata fields. Signed-off-by: Srikanth Yalavarthi --- drivers/ml/cnxk/

[PATCH v1 1/3] ml/cnxk: split metadata fields into sections

Split metadata into header, model sections, weights & bias, input / output and data sections. This is a preparatory step to introduce v2301 of model metadata. Signed-off-by: Srikanth Yalavarthi --- drivers/ml/cnxk/cn10k_ml_model.c | 26 +- drivers/ml/cnxk/cn10k_ml_model.h | 487

[PATCH v1 0/3] Add support for 32 I/O per model

This patch series adds support for 32 inputs / outputs per each model. Changes required to enable the required support include: 1. Splitiing model metadata fields into structures. 2. Update model metadata to v2301 which supports 32 I/O. 3. Update ML driver code to support metadata v2301 . Srikan

[PATCH v1] ml/cnxk: fix xstat type names in documentation

Fix incorrect type names for xstats in ML cnxk driver documentation. Fixes: 4ff4ab8e1a20 ("ml/cnxk: support extended statistics") Signed-off-by: Srikanth Yalavarthi --- doc/guides/mldevs/cnxk.rst | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/doc/guides/mldevs/cnxk.rs

[PATCH v2] app/mldev: fix code formatting and typos

Updated ML application source files to have uniform code formatting style across. Remove extra blank lines. Fix typos in application help. Fixes: 8cb22a545447 ("app/mldev: fix debug build") Fixes: da6793390596 ("app/mldev: support inference validation") Fixes: c0e871657d6a ("app/mldev: support que

[PATCH v2] app/mldev: add internal function for file read

Added internal function to read model, input and reference files with required error checks. This change fixes the unchecked return value and improper use of negative value issues reported by coverity scan for file read operations. Coverity issue: 383742, 383743 Fixes: f6661e6d9a3a ("app/mldev: va

Re: [EXT] [PATCH] crypto/uadk: set queue pair in dev_configure

Hi, Akhil On Thu, 20 Apr 2023 at 15:20, Akhil Goyal wrote: > > > By default, uadk only alloc two queues for each algorithm, which > > will impact performance. > > Set queue pair number as required in dev_configure. > > The default max queue pair number is 8, which can be modified > > via para: ma

[RFC PATCH 1/1] eventdev: introduce ML event adapter library

Introduce event ML adapter APIs. This patch provides information on adapter modes and usage. Application can use this event adapter interface to transfer packets between ML device and event device. Signed-off-by: Srikanth Yalavarthi --- MAINTAINERS |6 + con

[RFC PATCH 0/1] Introduce Event ML Adapter

Machine learning event adapter library == DPDK Eventdev library provides event driven programming model with features to schedule events. ML Device library provides an interface to ML poll mode drivers that support Machine Learning inference operations. Event

RE: [RFC] ring: improve ring performance with C11 atomics

> From: Wathsala Vithanage [mailto:wathsala.vithan...@arm.com] > Sent: Friday, 21 April 2023 21.17 > > Tail load in __rte_ring_move_cons_head and __rte_ring_move_prod_head > can be changed to __ATOMIC_RELAXED from __ATOMIC_ACQUIRE. > Because to calculate the addresses of the dequeue > elements __r