Update nvml.py to CUDA 11; smi.py DeviceQuery to R460 (#33)

* Updated for enums and structs for CUDA 9, 10, 10.1, 10.2, 11

* Added Dockerfile for pytest coverage

* adding nvml header files a requirement docs for implementation

* Added bindings for some functions added in CUDA 11.0

* Added support for more MIG functions

* Merged nvml.py and smi.py for CUDA 11 and R460 driver

* Updated for CUDA 11

* Delete nvml_cuda100.h

* Delete nvml_cuda110.h

* Delete nvml_cuda92.h

* Delete nvml_cuda101.h

* Delete nvml_cuda102.h

* Delete nvml_cuda90.h

* Delete nvml_cuda91.h

Co-authored-by: Travis Hester <[email protected]>

Author: kenhester

LICENSE.txt

@@ -1,4 +1,4 @@
-Copyright (c) 2011-2020, NVIDIA Corporation.
+Copyright (c) 2011-2021, NVIDIA Corporation.
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

README.md

@@ -47,6 +47,12 @@ nvsmi = nvidia_smi.getInstance()
 nvsmi.DeviceQuery('memory.free, memory.total')
 ```
 
+```python
+from pynvml.smi import nvidia_smi
+nvsmi = nvidia_smi.getInstance()
+print(nvsmi.DeviceQuery('--help-query-gpu'), end='\n')
+```
+
 Functions
 ---------
 Python methods wrap NVML functions, implemented in a C shared library.

docker/Dockerfile

@@ -0,0 +1,39 @@
+##############################################################
+# This Dockerfile contains the additional NVIDIA compilers, 
+# libraries, and plugins to enable OpenACC and NVIDIA GPU 
+# acceleration of Devito codes.
+#
+# BUILD: docker build --network=host --file docker/Dockerfile --tag pynvml .
+# RUN: docker run --gpus all --rm -it -p 8888:8888 -p 8787:8787 -p 8786:8786 pynvml
+##############################################################
+FROM python:3.6
+
+ENV DEBIAN_FRONTEND noninteractive 
+
+ADD ./requirements.txt /app/requirements.txt
+
+RUN pip install --no-cache-dir --upgrade pip && \
+    pip install --no-cache-dir -r /app/requirements.txt && \
+    rm -rf ~/.cache/pip
+
+ADD ./pynvml /app/pynvml
+ADD ./setup.py /app/
+ADD ./setup.cfg /app/
+ADD ./README.md /app/
+ADD ./PKG-INFO /app/
+ADD ./MANIFEST.in /app/
+ADD ./help_query_gpu.txt /app/
+ADD docker/entrypoint.sh /docker-entrypoint.sh
+
+RUN chmod +x  /docker-entrypoint.sh 
+
+## Create App user 
+# Set the home directory to our app user's home.
+ENV HOME=/app
+ENV APP_HOME=/app
+
+WORKDIR /app
+
+EXPOSE 8888
+ENTRYPOINT ["/docker-entrypoint.sh"]
+CMD ["bash"]

docker/entrypoint.sh

@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+
+find /app -type f -name '*.pyc' -delete
+
+export PATH=/venv/bin:$PATH
+export PYTHONPATH=$PYTHONPATH:/app
+
+exec "$@"

help_query_gpu.txt

@@ -107,7 +107,7 @@ The GOM currently in use.
 The GOM that will be used on the next reboot.
 
 "fan.speed" or NVSMI_FAN_SPEED
-The fan speed value is the percent of maximum speed that the device's fan is currently intended to run at. It ranges from 0 to 100 %. Note: The reported speed is the intended fan speed. If the fan is physically blocked and unable to spin, this output will not match the actual fan speed. Many parts do not report fan speeds because they rely on cooling via fans in the surrounding enclosure.
+The fan speed value is the percent of the product's maximum noise tolerance fan speed that the device's fan is currently intended to run at. This value may exceed 100% in certain cases. Note: The reported speed is the intended fan speed. If the fan is physically blocked and unable to spin, this output will not match the actual fan speed. Many parts do not report fan speeds because they rely on cooling via fans in the surrounding enclosure.
 
 "pstate" or NVSMI_PSTATE
 The current performance state for the GPU. States range from P0 (maximum performance) to P12 (minimum performance).
@@ -210,6 +210,9 @@ NVIDIA GPUs can provide error counts for various types of ECC errors. Some ECC e
 "ecc.errors.corrected.volatile.device_memory" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_DEV_MEM
 Errors detected in global device memory.
 
+"ecc.errors.corrected.volatile.dram" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_DRAM
+Errors detected in global device memory.
+
 "ecc.errors.corrected.volatile.register_file" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_REGFILE
 Errors detected in register file memory.
 
@@ -222,12 +225,21 @@ Errors detected in the L2 cache.
 "ecc.errors.corrected.volatile.texture_memory" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_TEXTURE
 Parity errors detected in texture memory.
 
+"ecc.errors.corrected.volatile.cbu" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_CBU
+Parity errors detected in CBU.
+
+"ecc.errors.corrected.volatile.sram" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_SRAM
+Errors detected in global SRAMs.
+
 "ecc.errors.corrected.volatile.total" or NVSMI_ECC_ERROR_CORRECTED_VOLATILE_TOTAL
-Total errors detected across entire chip. Sum of device_memory, register_file, l1_cache, l2_cache and texture_memory.
+Total errors detected across entire chip.
 
 "ecc.errors.corrected.aggregate.device_memory" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_DEV_MEM
 Errors detected in global device memory.
 
+"ecc.errors.corrected.aggregate.dram" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_DRAM
+Errors detected in global device memory.
+
 "ecc.errors.corrected.aggregate.register_file" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_REGFILE
 Errors detected in register file memory.
 
@@ -240,12 +252,21 @@ Errors detected in the L2 cache.
 "ecc.errors.corrected.aggregate.texture_memory" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_TEXTURE
 Parity errors detected in texture memory.
 
+"ecc.errors.corrected.aggregate.cbu" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_CBU
+Parity errors detected in CBU.
+
+"ecc.errors.corrected.aggregate.sram" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_SRAM
+Errors detected in global SRAMs.
+
 "ecc.errors.corrected.aggregate.total" or NVSMI_ECC_ERROR_CORRECTED_AGGREGATE_TOTAL
 Total errors detected across entire chip. Sum of device_memory, register_file, l1_cache, l2_cache and texture_memory.
 
 "ecc.errors.uncorrected.volatile.device_memory" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_DEV_MEM
 Errors detected in global device memory.
 
+"ecc.errors.uncorrected.volatile.dram" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_DRAM
+Errors detected in global device memory.
+
 "ecc.errors.uncorrected.volatile.register_file" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_REGFILE
 Errors detected in register file memory.
 
@@ -258,12 +279,21 @@ Errors detected in the L2 cache.
 "ecc.errors.uncorrected.volatile.texture_memory" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_TEXTURE
 Parity errors detected in texture memory.
 
+"ecc.errors.uncorrected.volatile.cbu" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_CBU
+Parity errors detected in CBU.
+
+"ecc.errors.uncorrected.volatile.sram" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_SRAM
+Errors detected in global SRAMs.
+
 "ecc.errors.uncorrected.volatile.total" or NVSMI_ECC_ERROR_UNCORRECTED_VOLATILE_TOTAL
-Total errors detected across entire chip. Sum of device_memory, register_file, l1_cache, l2_cache and texture_memory.
+Total errors detected across entire chip. 
 
 "ecc.errors.uncorrected.aggregate.device_memory" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_DEV_MEM
 Errors detected in global device memory.
 
+"ecc.errors.uncorrected.aggregate.dram"  or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_DRAM
+Errors detected in global device memory.
+
 "ecc.errors.uncorrected.aggregate.register_file" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_REGFILE
 Errors detected in register file memory.
 
@@ -276,8 +306,14 @@ Errors detected in the L2 cache.
 "ecc.errors.uncorrected.aggregate.texture_memory" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_TEXTURE
 Parity errors detected in texture memory.
 
+"ecc.errors.uncorrected.aggregate.cbu" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_CBU
+Parity errors detected in CBU.
+
+"ecc.errors.uncorrected.aggregate.sram" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_SRAM
+Errors detected in global SRAMs.
+
 "ecc.errors.uncorrected.aggregate.total" or NVSMI_ECC_ERROR_UNCORRECTED_AGGREGATE_TOTAL
-Total errors detected across entire chip. Sum of device_memory, register_file, l1_cache, l2_cache and texture_memory.
+Total errors detected across entire chip. 
 
 Section about retired_pages properties
 NVIDIA GPUs can retire pages of GPU device memory when they become unreliable. This can happen when multiple single bit ECC errors occur for the same page, or on a double bit ECC error. When a page is retired, the NVIDIA driver will hide it such that no driver, or application memory allocations can access it.
@@ -360,6 +396,17 @@ Maximum frequency of SM (Streaming Multiprocessor) clock.
 "clocks.max.memory" or "clocks.max.mem" or NVSMI_CLOCKS_MEMORY_MAX
 Maximum frequency of memory clock.
 
+Section about mig.mode properties
+A flag that indicates whether MIG mode is enabled. May be either "Enabled" or "Disabled". Changes to MIG mode require a GPU reset.
+
+"mig.mode.current" or NVSMI_MIG_MODE_CURRENT
+The MIG mode that the GPU is currently operating under.
+
+"mig.mode.pending" or NVSMI_MIG_MODE_PENDING
+The MIG mode that the GPU will operate under after reset.
+
+Section about commands to return additional properties
+
 "supported-clocks" or NVSMI_CLOCKS_SUPPORTED
 List of supported clocks.