From: Srikanth Yalavarthi <syalavarthi@marvell.com>
To: Srikanth Yalavarthi <syalavarthi@marvell.com>
Cc: <dev@dpdk.org>, <aprabhu@marvell.com>, <sshankarnara@marvell.com>,
<ptakkar@marvell.com>
Subject: [PATCH 3/3] ml/cnxk: add support for additional integer types
Date: Sun, 7 Jan 2024 07:28:12 -0800 [thread overview]
Message-ID: <20240107152813.2668-4-syalavarthi@marvell.com> (raw)
In-Reply-To: <20240107152813.2668-1-syalavarthi@marvell.com>
Added support quantization and dequantization of 32-bit
and 64-bit integer types.
Signed-off-by: Srikanth Yalavarthi <syalavarthi@marvell.com>
---
drivers/ml/cnxk/cnxk_ml_io.c | 24 ++++++++++++++++++++++++
drivers/ml/cnxk/mvtvm_ml_model.c | 4 ++++
2 files changed, 28 insertions(+)
diff --git a/drivers/ml/cnxk/cnxk_ml_io.c b/drivers/ml/cnxk/cnxk_ml_io.c
index c78009ab0cd..4b0adc2ae47 100644
--- a/drivers/ml/cnxk/cnxk_ml_io.c
+++ b/drivers/ml/cnxk/cnxk_ml_io.c
@@ -40,6 +40,18 @@ cnxk_ml_io_quantize_single(struct cnxk_ml_io *input, uint8_t *dbuffer, uint8_t *
case RTE_ML_IO_TYPE_UINT16:
ret = rte_ml_io_float32_to_uint16(qscale, nb_elements, dbuffer, qbuffer);
break;
+ case RTE_ML_IO_TYPE_INT32:
+ ret = rte_ml_io_float32_to_int32(qscale, nb_elements, dbuffer, qbuffer);
+ break;
+ case RTE_ML_IO_TYPE_UINT32:
+ ret = rte_ml_io_float32_to_uint32(qscale, nb_elements, dbuffer, qbuffer);
+ break;
+ case RTE_ML_IO_TYPE_INT64:
+ ret = rte_ml_io_float32_to_int64(qscale, nb_elements, dbuffer, qbuffer);
+ break;
+ case RTE_ML_IO_TYPE_UINT64:
+ ret = rte_ml_io_float32_to_uint64(qscale, nb_elements, dbuffer, qbuffer);
+ break;
case RTE_ML_IO_TYPE_FP16:
ret = rte_ml_io_float32_to_float16(nb_elements, dbuffer, qbuffer);
break;
@@ -82,6 +94,18 @@ cnxk_ml_io_dequantize_single(struct cnxk_ml_io *output, uint8_t *qbuffer, uint8_
case RTE_ML_IO_TYPE_UINT16:
ret = rte_ml_io_uint16_to_float32(dscale, nb_elements, qbuffer, dbuffer);
break;
+ case RTE_ML_IO_TYPE_INT32:
+ ret = rte_ml_io_int32_to_float32(dscale, nb_elements, qbuffer, dbuffer);
+ break;
+ case RTE_ML_IO_TYPE_UINT32:
+ ret = rte_ml_io_uint32_to_float32(dscale, nb_elements, qbuffer, dbuffer);
+ break;
+ case RTE_ML_IO_TYPE_INT64:
+ ret = rte_ml_io_int64_to_float32(dscale, nb_elements, qbuffer, dbuffer);
+ break;
+ case RTE_ML_IO_TYPE_UINT64:
+ ret = rte_ml_io_uint64_to_float32(dscale, nb_elements, qbuffer, dbuffer);
+ break;
case RTE_ML_IO_TYPE_FP16:
ret = rte_ml_io_float16_to_float32(nb_elements, qbuffer, dbuffer);
break;
diff --git a/drivers/ml/cnxk/mvtvm_ml_model.c b/drivers/ml/cnxk/mvtvm_ml_model.c
index 0dbe08e9889..e3234ae4422 100644
--- a/drivers/ml/cnxk/mvtvm_ml_model.c
+++ b/drivers/ml/cnxk/mvtvm_ml_model.c
@@ -150,6 +150,8 @@ mvtvm_ml_io_type_map(DLDataType dltype)
return RTE_ML_IO_TYPE_INT16;
else if (dltype.bits == 32)
return RTE_ML_IO_TYPE_INT32;
+ else if (dltype.bits == 64)
+ return RTE_ML_IO_TYPE_INT64;
break;
case kDLUInt:
if (dltype.bits == 8)
@@ -158,6 +160,8 @@ mvtvm_ml_io_type_map(DLDataType dltype)
return RTE_ML_IO_TYPE_UINT16;
else if (dltype.bits == 32)
return RTE_ML_IO_TYPE_UINT32;
+ else if (dltype.bits == 64)
+ return RTE_ML_IO_TYPE_UINT64;
break;
case kDLFloat:
if (dltype.bits == 8)
--
2.42.0
next prev parent reply other threads:[~2024-01-07 15:28 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-07 15:28 [PATCH 0/3] add support for additional data types Srikanth Yalavarthi
2024-01-07 15:28 ` [PATCH 1/3] mldev: add conversion routines for 32-bit integers Srikanth Yalavarthi
2024-01-07 15:28 ` [PATCH 2/3] mldev: add support for 64-integer data type Srikanth Yalavarthi
2024-01-07 15:28 ` Srikanth Yalavarthi [this message]
2024-02-18 22:14 ` [PATCH 0/3] add support for additional data types Thomas Monjalon
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=20240107152813.2668-4-syalavarthi@marvell.com \
--to=syalavarthi@marvell.com \
--cc=aprabhu@marvell.com \
--cc=dev@dpdk.org \
--cc=ptakkar@marvell.com \
--cc=sshankarnara@marvell.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;
as well as URLs for NNTP newsgroup(s).