compute_unit 0.1.0 → 0.3.2

Sign up to get free protection for your applications and to get access to all the features.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 5caf721411458c1307f2714ba330c737ddd5987433a51d7753b6004649a92429
4
- data.tar.gz: ef3bacd1c4662e1ffe33dafad9c48f44c6fc3cefe161098957cc476e7ecac78d
3
+ metadata.gz: d3fe6b765613752707aca6037a41159d6a3d1d0f33638b3ab5b20ed7d693d844
4
+ data.tar.gz: bbc7c51a56dcca030098341bfe587ff4263d28d44859b13e9ad2efffb5614a46
5
5
  SHA512:
6
- metadata.gz: 73b107433ce4ffcb3d6a5d35a8711cf39489ed111ccc368cb66d5f788f36690382242a3437c0bd6d40147c2ca727856d2fffd48b8b5edbaa8f4dd22b5ab228da
7
- data.tar.gz: 2debc1c284e7200efd00ca0e064d3b48f55a935d55409cc77f33db2db9344efabcdaadc3cd2bfd909174e58ba38727ed4ed845dad0395a19e08dc5ee0f54a428
6
+ metadata.gz: 4653e9bd5a0c9da77403b6afa8fd3cd9b07c95399c074b140b82201742ef50fb80cbfd4221c792aee17d2b60547f96db2265d2a4d6797b3558bab7ffb9e97416
7
+ data.tar.gz: ef104cce6547acdc6390e5d1ab315b11eed5aaa5dcf548198f531b2bca30bec58220f23d062f2e7e7076bfc137056980fc35f34517577bdb2331e77b5d8ad840
@@ -1,3 +1,31 @@
1
1
  # Compute Unit Changelog
2
2
 
3
+ ## Unrleased
4
+ ## 0.3.2
5
+ * Fix bug with default database not being created
6
+
7
+ ## 0.3.1
8
+ Released 9/24/2020
9
+
10
+ * Use the default system pcidb file instead of throwing error
11
+ * Return default data if no hmon value exist
12
+
13
+ ## 0.3.0
14
+ Released 6/16/2020
15
+
16
+ * Add name method to the cpu class
17
+ * Use constant for proc path
18
+ * Use debug when kernel file doesn't exist
19
+
20
+ ## 0.2.1
21
+ Released 6/6/2020
22
+
23
+ * update rake and other gems
24
+
25
+ ## 0.2.0
26
+ Released 6/5/2020
27
+ * Added more cpu methods for various POI.
28
+ * Ensure constants exist when side loaded
29
+ * Add descendent discovery mechanism to GPU class
30
+
3
31
  ## initial version 0.1.0
data/Gemfile CHANGED
@@ -2,17 +2,11 @@
2
2
 
3
3
  source 'https://rubygems.org'
4
4
 
5
-
6
- gem 'opencl_ruby_ffi', '~> 1.3.4'
7
- gem 'compute_unit', path: '.'
8
-
9
5
  group :test do
10
6
  gem 'pry'
11
7
  gem 'rubocop'
12
8
  gem 'rubocop-rspec'
13
9
  gem 'ruby-prof'
14
10
  gem 'simplecov'
15
- gem 'rake', '~> 10.0'
16
- gem 'rspec', '~> 3.0'
17
- gem 'bundler', '~> 2.0'
18
11
  end
12
+ gemspec
data/README.md CHANGED
@@ -1,9 +1,9 @@
1
1
  # Compute Unit Ruby Library
2
- The compute unit library helps you write future abstractions and tooling for crypto mining, machine learning or just general purpose usage. This library was extracted out of the crypto mining tool [Crossbelt](https://crossbelt.blockops.party) to help spur more tooling to ruby.
2
+ The compute unit library helps you write future abstractions and tooling for crypto mining, machine learning or just general purpose. This library was extracted out of the crypto mining tool [Crossbelt](https://crossbelt.blockops.party) to help spur more tooling for ruby.
3
3
 
4
4
  Uses the sysfs filesystem to quickly read and write compute unit data. More information about sysfs can be found [here](https://www.kernel.org/doc/Documentation/filesystems/sysfs-pci.txt)
5
5
 
6
- This library will:
6
+ This library has the ability to:
7
7
 
8
8
  * Detect all compute devices such as cpu, gpu, and anything else we add to this library.
9
9
  * Provide the ability to control GPU fans.
@@ -11,6 +11,7 @@ This library will:
11
11
  * Change or view voltage, core clock, memory clock and other tunning parameters used for overclocking.
12
12
  * Provide inventory data of all the devices with make, model and other important information.
13
13
  * Provide realtime status information of various data points. Useful for real-time tunning.
14
+ * Interact with the compute device via ruby.
14
15
 
15
16
  This library was open sourced to allow further refinement and usage throughout the ruby community.
16
17
  Since machine learning is not popular on ruby due to device and program support, I hope this gem will
@@ -34,12 +35,133 @@ Or install it yourself as:
34
35
 
35
36
  ## Usage
36
37
 
38
+ Find all compute devices
37
39
  ```
38
40
  require 'compute_unit'
39
41
  devices = ComputeUnit.find_all
40
42
 
41
43
  ```
42
44
 
45
+ Find just GPU devices and list the voltage table
46
+
47
+ ```ruby
48
+ require 'compute_unit/gpu'
49
+
50
+ gpu = ComputeUnit::Gpu.find_all.first
51
+ gpu.voltage_table
52
+ => [{:pstate=>0, :clk=>852, :volt=>800, :type=>:sclk},
53
+ {:pstate=>1, :clk=>997, :volt=>860, :type=>:sclk},
54
+ {:pstate=>2, :clk=>1084, :volt=>860, :type=>:sclk},
55
+ {:pstate=>3, :clk=>1138, :volt=>860, :type=>:sclk},
56
+ {:pstate=>4, :clk=>1200, :volt=>860, :type=>:sclk},
57
+ {:pstate=>5, :clk=>1401, :volt=>860, :type=>:sclk},
58
+ {:pstate=>6, :clk=>1536, :volt=>860, :type=>:sclk},
59
+ {:pstate=>7, :clk=>1630, :volt=>1200, :type=>:sclk}]
60
+
61
+ gpu.amdgpu_pm_info
62
+ => {:mclk=>{:value=>"1020", :unit=>"MHz"},
63
+ :sclk=>{:value=>"994", :unit=>"MHz"},
64
+ :pstate_sclk=>{:value=>"1138", :unit=>"MHz"},
65
+ :pstate_mclk=>{:value=>"800", :unit=>"MHz"},
66
+ :vddgfx=>{:value=>"862", :unit=>"mV"},
67
+ :average_gpu=>{:value=>"121.0", :unit=>"W"},
68
+ :temperature=>{:value=>"57", :unit=>"C"},
69
+ :load=>{:value=>"100", :unit=>"%"}}
70
+
71
+ ```
72
+
73
+ ## Example CLI command
74
+ There is a `list_computes` command that will enumerate all the compute devices found on the system.
75
+
76
+ ```ruby
77
+ list_computes
78
+ [
79
+ {
80
+ "uuid": "GPU0",
81
+ "gpuId": "GPU0",
82
+ "syspath": "/sys/bus/pci/devices/0000:03:00.0",
83
+ "pciLoc": "0000:03:00.0",
84
+ "name": "RX Vega56",
85
+ "bios": "113-D0500100-102",
86
+ "subType": "amdgpu",
87
+ "make": "AMD",
88
+ "model": "RX Vega56",
89
+ "vendor": "AMD",
90
+ "power": 121,
91
+ "utilization": 100,
92
+ "temperature": 57,
93
+ "status": 0,
94
+ "pstate": -1,
95
+ "fanSpeed": 2946,
96
+ "type": "GPU",
97
+ "maxTemp": null,
98
+ "mem": 1020,
99
+ "cor": 997,
100
+ "vlt": 862,
101
+ "mem_temp": 75,
102
+ "maxFan": null,
103
+ "dpm": null,
104
+ "vddci": null,
105
+ "maxPower": null,
106
+ "ocProfile": null,
107
+ "opencl_enabled": false
108
+ },
109
+ {
110
+ "chip_make": "GenuineIntel",
111
+ "make": "GenuineIntel",
112
+ "model": "Intel(R) Core(TM) i3-7100 CPU @ 3.90GHz",
113
+ "device_id": "590f",
114
+ "vendor_id": "8086",
115
+ "subsystem_device_id": "7a59",
116
+ "subsystem_vendor_id": "1462",
117
+ "device_class": "060000",
118
+ "temp": 33,
119
+ "minFreqMhz": 800,
120
+ "maxFreqMhz": 3900,
121
+ "currentFreqMhz": 3600,
122
+ "numCores": 2,
123
+ "numThreads": 2,
124
+ "nproc": 4,
125
+ "temps": {
126
+ "package_id_0": 33,
127
+ "core_0": 26,
128
+ "core_1": 33
129
+ }
130
+ }
131
+ ]
132
+ ```
133
+
134
+ ### Subclassing a compute object
135
+ Instead of monkey patching this code you can simply subclass the objects. The find_all methods are intelligent enough to locate your subclass by searching the Objectspace for all decendants of `ComputeUnit::ComputeBase`. This makes it dead simple to add new functionality without having to wrap any code. This gives you the ability to include new modules or override existing methods. Just make sure you load your class before using the find_all method.
136
+
137
+
138
+
139
+ ```ruby
140
+ # coolproject/vegagpu.rb
141
+ require 'cool_project/vega_gpu'
142
+ module CoolProject
143
+ class VegaGpu < ::ComputeUnit::Gpu::AmdGpu
144
+ include CoolProject::Metrics
145
+
146
+ def initialize(device_path, opts)
147
+ super(device_path, opts)
148
+ @is_vega = true
149
+ end
150
+ end
151
+ end
152
+
153
+ ```
154
+
155
+ Once you have your new class created all you need to do is load the class before running find all.
156
+
157
+ ```ruby
158
+ require 'compute_unit'
159
+ require 'coolproject/vegagpu'
160
+ require 'yaml'
161
+
162
+ puts ComputeUnit.find_all.to_yaml
163
+
164
+ ```
43
165
 
44
166
  ## Make and Model information
45
167
  This library uses the pci database maintained here https://pci-ids.ucw.cz/ which is also used by `lspci` and other linux tools. If your device is giving the wrong info the problem is likely due to the database being incorrect.
@@ -51,9 +173,25 @@ If you want to refresh the database every time with finding all the devices you
51
173
  Or you can also run the simple cli tool. `update_pcidb` that comes with this library.
52
174
 
53
175
  ## Kernel support
54
- There will likely be changes with reading and writing information to sysfs as new kerneles are released. This library has been tested against.
176
+ There will likely be changes with reading and writing information to sysfs as new kernels are released. This library has been tested against.
55
177
 
56
178
  * 4.15 - 4.20
179
+ * 5.x
180
+
181
+ ## OpenCL Usage
182
+ The opencl_ruby_ffi gem is optionally used to gather additional info about the GPU. This can sometimes result in better model info.
183
+
184
+ There is some caution when using OpenCL. If a GPU is dead, OpenCL tends to hang/freeze thus forcing a reboot. This is not a bug in the compute_unit or opencl_ruby_ffi gem but in the OpenCL library instead.
185
+
186
+ Getting data from OpenCL also takes a long time! So if you have multiple GPUs in your system this could takes multiple seconds to return data. Because of this we cache the responses for the next query. This response is cached until the system checksum changes. This checksum is a special calculate of all the devices present in the system. So if you ever added or changed the equipment in the system, a new OpenCL query would ensue.
187
+
188
+ To enable OpenCL you just need to pass a boolean value to any of the find_all methods:
189
+
190
+ ```ruby
191
+ require 'compute_unit'
192
+ devices = ComputeUnit.find_all(true)
193
+
194
+ ```
57
195
 
58
196
  ## Development
59
197
 
@@ -67,7 +205,7 @@ Bug reports and pull requests are welcome on GitHub at https://github.com/logicm
67
205
 
68
206
  ## License
69
207
 
70
- The gem is available as open source under the terms of the [MIT License](https://opensource.org/licenses/MIT).
208
+ The gem is available as open source under the terms of the [MIT License](https://opensource.org/licenses/MIT.
71
209
 
72
210
  ## Code of Conduct
73
211
 
@@ -13,9 +13,9 @@ Gem::Specification.new do |spec|
13
13
  spec.summary = 'A ruby library for compute unit devices'
14
14
  spec.description = <<~EOF
15
15
 
16
- A ruby library that searches uses the linux sysfs file system for compute unit devices such as
17
- CPUS, GPUs and other ASIC compute devices. Allows programmatic access collect real time metrics from the kernel or relatated driver toolchain.
18
- Is meant to be used as a toolchain to future build tooling on. This library also makes use of opencl toolchain and requires
16
+ A ruby library that searches the linux sysfs file system for compute unit devices such as
17
+ CPUS, GPUs and other ASIC compute devices. Allows programmatic access to collect real time metrics from the kernel or relatated driver toolchain.
18
+ Is meant to be used as a toolchain for future tools. This library also makes use of opencl library and requires
19
19
  the opencl_ruby_ffi gem.
20
20
 
21
21
  EOF
@@ -25,7 +25,7 @@ Gem::Specification.new do |spec|
25
25
 
26
26
  spec.metadata['homepage_uri'] = spec.homepage
27
27
  spec.metadata['source_code_uri'] = spec.homepage
28
- spec.metadata['changelog_uri'] = "#{spec.homepage}/CHANGELOG"
28
+ spec.metadata['changelog_uri'] = "#{spec.homepage}/-/blob/master/CHANGELOG.md"
29
29
 
30
30
  # Specify which files should be added to the gem when it is released.
31
31
  # The `git ls-files -z` loads the files in the RubyGem that have been added into git.
@@ -36,8 +36,8 @@ Gem::Specification.new do |spec|
36
36
  spec.executables = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
37
37
  spec.require_paths = ['lib']
38
38
 
39
- spec.add_dependency 'opencl_ruby_ffi', '~> 1.3.0'
40
- spec.add_development_dependency 'bundler', '~> 2.0'
41
- spec.add_development_dependency 'rake', '~> 10.0'
39
+ spec.add_dependency 'opencl_ruby_ffi', '~> 1.3.4'
40
+ spec.add_development_dependency 'bundler', '~> 2.1'
41
+ spec.add_development_dependency 'rake', '~> 13'
42
42
  spec.add_development_dependency 'rspec', '~> 3.0'
43
43
  end
@@ -11,23 +11,36 @@ module ComputeUnit
11
11
  SYS_DEVICE_PATH = File.join(SYSFS_PATH, 'bus/pci/devices')
12
12
  PCI_DATABASE_PATH = File.join(File.dirname(__dir__), 'pci.ids')
13
13
  PCI_DATABASE_URL = 'http://pci-ids.ucw.cz/v2.2/pci.ids'
14
+ DEFAULT_PCIDB_PATH = '/usr/share/hwdata/pci.ids'
14
15
 
16
+ # @param use_opencl [Boolean]
17
+ # @return [Array] - return a list of compute units
15
18
  def self.find_all(use_opencl = false)
16
19
  require 'compute_unit/gpu'
17
20
  require 'compute_unit/cpu'
18
21
  require 'compute_unit/asic'
19
- raise Exceptions::InvalidPCIDatabase.new('Run: ComputeUnit.refresh_pci_database') unless File.exist?(PCI_DATABASE_PATH)
22
+
23
+ copy_default_database unless File.exist?(PCI_DATABASE_PATH)
20
24
 
21
25
  ComputeUnit::Gpu.find_all(use_opencl) +
22
26
  ComputeUnit::Cpu.find_all +
23
27
  ComputeUnit::Asic.find_all
24
28
  end
25
29
 
30
+ # copies the default pci database from linux filesystem over to the gem path
31
+ def self.copy_default_database
32
+ File.cp(DEFAULT_PCIDB_PATH, PCI_DATABASE_PATH)
33
+ end
34
+
35
+ # get a fresh copy of the database and then use find_all
36
+ # @param use_opencl [Boolean]
37
+ # @return [Array] - return a list of compute units
26
38
  def self.find_all_with_database(use_opencl = false)
27
39
  refresh_pci_database
28
40
  find_all(use_opencl)
29
41
  end
30
42
 
43
+ # downloads the pci database
31
44
  def self.refresh_pci_database
32
45
  ComputeUnit::Utils.check_for_root
33
46
  require 'net/http'
@@ -9,6 +9,25 @@ module ComputeUnit
9
9
  []
10
10
  end
11
11
 
12
+ def self.create_from_path(device_path, index)
13
+ opts = {
14
+ device_class_id: device_class(device_path),
15
+ device_id: device(device_path),
16
+ device_vendor_id: device_vendor(device_path),
17
+ subsystem_vendor_id: subsystem_vendor(device_path),
18
+ subsystem_device_id: subsystem_device(device_path),
19
+ index: index
20
+ }
21
+ new(device_path, opts)
22
+ end
23
+
24
+ def self.find_all
25
+ return []
26
+ devices.sort.map.with_index do |device_path, index|
27
+ create_from_path(device_path, index)
28
+ end
29
+ end
30
+
12
31
  # /sys/devices/platform $ more coretemp.0/hwmon/hwmon1/temp1_input / 1000
13
32
  end
14
33
  end
@@ -16,6 +16,16 @@ module ComputeUnit
16
16
  # timeout value
17
17
  CACHE_TIMEOUT = 30
18
18
 
19
+ # @return [Array] - find all the decendants of thyself
20
+ def self.compute_classes
21
+ ObjectSpace.each_object(Class).select do |klass|
22
+ # <Class:#<Crossbelt::ComputeUnit::NvidiaGpu:0x00007fddc5c02a10>>
23
+ # We have to filter out these kinds of Ojbects as they don't respond to :new
24
+ # without a singleton error.
25
+ klass < self && !klass.to_s.include?('Class')
26
+ end
27
+ end
28
+
19
29
  # @param value [Float] a value to offset the power calculation, either a whole number, or decimal
20
30
  # @return [Integer] the value set as the offset
21
31
  def power_offset=(value)
@@ -7,12 +7,86 @@ module ComputeUnit
7
7
  DEVICE_CLASS = '060000'
8
8
  DEVICE_CLASS_NAME = 'CPU'
9
9
  # @return [Array] - returns a list of device paths of all devices considered for display
10
+ alias name model
11
+
10
12
  def self.devices
11
13
  ComputeUnit::ComputeBase.devices.find_all do |device|
12
14
  ComputeUnit::Device.device_class(device) == DEVICE_CLASS
13
15
  end
14
16
  end
15
17
 
18
+ def model
19
+ raw_cpu_data[:model_name]
20
+ end
21
+
22
+ def make
23
+ raw_cpu_data[:vendor_id]
24
+ end
25
+
26
+ def num_cores
27
+ raw_cpu_data[:cores_per_socket].to_i
28
+ end
29
+
30
+ def num_threads
31
+ raw_cpu_data[:threads_per_core].to_i
32
+ end
33
+
34
+ def nproc
35
+ raw_cpu_data[:cpus].to_i
36
+ end
37
+
38
+ def current_freq_mhz
39
+ raw_cpu_data[:cpu_mhz].to_f.round(0)
40
+ end
41
+
42
+ def max_freq_mhz
43
+ raw_cpu_data[:cpu_max_mhz].to_f.round(0)
44
+ end
45
+
46
+ def min_freq_mhz
47
+ raw_cpu_data[:cpu_min_mhz].to_f.round(0)
48
+ end
49
+
50
+ def base_hwmon_path
51
+ File.join(ComputeUnit::SYSFS_PATH, 'devices/platform/coretemp.0/hwmon')
52
+ end
53
+
54
+ # @return [Hash] - a hash of temp readings and their labels
55
+ # @example temps => {:core_0=>31, :core_1=>31, :package_id_0=>27}
56
+ def temps
57
+ Dir.glob(File.join(hwmon_path, 'temp*_label')).each_with_object({}) do |label_file, acc|
58
+ temp_file = label_file.sub('label', 'input')
59
+ label = normalize_name(read_file(label_file))
60
+ reading = read_file(temp_file).to_f.round(0) / 1000
61
+ acc[label] = reading
62
+ acc
63
+ end
64
+ end
65
+
66
+ # @return [Integer] - the temperature of the cpu package in Celsius
67
+ def temp
68
+ read_hwmon_data('temp1_input', 0).to_f.round(0) / 1000
69
+ end
70
+
71
+ def metrics
72
+ {
73
+ temp: temp,
74
+ minFreqMhz: min_freq_mhz,
75
+ maxFreqMhz: max_freq_mhz,
76
+ currentFreqMhz: current_freq_mhz,
77
+ model: model,
78
+ make: make,
79
+ numCores: num_cores,
80
+ numThreads: num_threads,
81
+ nproc: nproc,
82
+ temps: temps
83
+ }
84
+ end
85
+
86
+ def to_h
87
+ super.merge(metrics)
88
+ end
89
+
16
90
  def self.create_from_path(device_path, index)
17
91
  opts = {
18
92
  device_class_id: device_class(device_path),
@@ -26,11 +100,51 @@ module ComputeUnit
26
100
  end
27
101
 
28
102
  def self.find_all
29
- return [] # not quite ready to show cpu data
30
103
  devices.sort.map.with_index do |device_path, index|
31
104
  create_from_path(device_path, index)
32
105
  end
33
106
  end
34
- # /sys/devices/platform $ more coretemp.0/hwmon/hwmon1/temp1_input / 1000
107
+
108
+ private
109
+
110
+ # @param value [String]
111
+ # @return [String]
112
+ # @summary downcases and transforms spaces into underscores
113
+ def normalize_name(value)
114
+ value.downcase.gsub(/\s/, '_').gsub('(s)', 's').to_sym
115
+ end
116
+
117
+ # @return [Hash] - a hash of cpu info returned from lscpu
118
+ # normalizes the key name by downcasing and removing spaces
119
+ # @example
120
+ # raw_cpu_data() =>
121
+ # {"architecture"=>"x86_64", "cpu_op-modes"=>"32-bit, 64-bit", "byte_order"=>"Little Endian",
122
+ # "cpus"=>"4", "on-line_cpus list"=>"0-3", "threads_per core"=>"2", "cores_per socket"=>"2",
123
+ # "sockets"=>"1", "numa_nodes"=>"1", "vendor_id"=>"GenuineIntel", "cpu_family"=>"6",
124
+ # "model"=>"158", "model_name"=>"Intel(R) Core(TM) i3-7100 CPU @ 3.90GHz", "stepping"=>"9",
125
+ # "cpu_mhz"=>"800.545", "cpu_max mhz"=>"3900.0000", "cpu_min mhz"=>"800.0000",
126
+ # "bogomips"=>"7799.87", "virtualization"=>"VT-x", "l1d_cache"=>"32K", "l1i_cache"=>"32K",
127
+ # "l2_cache"=>"256K", "l3_cache"=>"3072K", "numa_node0 cpus"=>"0-3",
128
+ # "flags"=>"fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36
129
+ # clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc
130
+ # art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni
131
+ # pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1
132
+ # sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm
133
+ # 3dnowprefetch cpuid_fault invpcid_single tpr_shadow vnmi flexpriority ept vpid ept_ad
134
+ # fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt
135
+ # xsaveopt xsavec xgetbv1 xsaves dtherm arat pln pts hwp hwp_notify hwp_act_window hwp_epp"
136
+ # }
137
+ def raw_cpu_data
138
+ @raw_cpu_data ||= begin
139
+ # turns each line into a hash of key value pairs
140
+ regex = /(?<name>[\s\w\-\(\)]+)\:\s+(?<value>[@\s\w\-\,\.\(\)]+)$/
141
+ `lscpu`.lines.each_with_object({}) do |line, acc|
142
+ m = line.match(regex)
143
+ name = normalize_name(m.named_captures['name'])
144
+ acc[name] = m.named_captures['value'].chomp if m
145
+ acc
146
+ end
147
+ end
148
+ end
35
149
  end
36
150
  end
@@ -1,16 +1,16 @@
1
1
  # frozen_string_literal: true
2
2
 
3
+ require 'compute_unit'
3
4
  require 'compute_unit/logger'
4
5
  require 'compute_unit/cache_store'
5
6
  require 'compute_unit/utils'
6
-
7
7
  # This file supports reading from sysfs
8
8
  # More information about sysfs can be found here - https://www.kernel.org/doc/Documentation/filesystems/sysfs-pci.txt
9
9
  module ComputeUnit
10
10
  class Device
11
11
  # We can supply a mock sysfs path in order to test with containers and other scenarios
12
12
  SYSFS_DEVICES_PATH = File.join(ComputeUnit::SYSFS_PATH, 'bus', 'pci', 'devices')
13
-
13
+ PROC_PATH = ENV['PROC_PATH'] || '/proc'
14
14
  attr_reader :device_class_id,
15
15
  :device_id,
16
16
  :device_vendor_id,
@@ -127,7 +127,10 @@ module ComputeUnit
127
127
 
128
128
  def read_file(path, default = nil)
129
129
  File.read(path).chomp
130
- rescue Errno::EINVAL, Errno::EPERM, Errno::ENOENT
130
+ rescue Errno::EINVAL, Errno::EPERM
131
+ default
132
+ rescue Errno::ENOENT
133
+ logger.debug("File #{path} does not exist, using defaults")
131
134
  default
132
135
  rescue Errno::EACCES
133
136
  logger.fatal('run this command as root or with sudo, using default value')
@@ -137,9 +140,9 @@ module ComputeUnit
137
140
  # @param item [String] - the name of the hwmon file to read from
138
141
  # @param default [Object] - the default value to return if the file is empty or not readable
139
142
  # @return [String] - the value of the item looked up
140
- def read_hwmon_data(item, _default = nil)
143
+ def read_hwmon_data(item, default = nil)
141
144
  path = File.join(hwmon_path, item)
142
- read_file(path)
145
+ read_file(path, default)
143
146
  end
144
147
 
145
148
  # @param item [String] - the name of the hwmon file to write to
@@ -268,7 +271,7 @@ module ComputeUnit
268
271
  logger.fatal(e.message)
269
272
  default
270
273
  rescue Errno::ENOENT
271
- logger.fatal("File #{path} does not exist")
274
+ logger.debug("File #{path} does not exist")
272
275
  default
273
276
  rescue Errno::EACCES
274
277
  logger.fatal('Run this command as root or with sudo')
@@ -286,7 +289,7 @@ module ComputeUnit
286
289
  rescue Errno::EINVAL, Errno::EPERM => e
287
290
  logger.fatal(e.message)
288
291
  rescue Errno::ENOENT
289
- logger.fatal("File #{path} does not exist")
292
+ logger.warn("File #{path} does not exist")
290
293
  rescue Errno::EACCES
291
294
  logger.fatal('Run this command as root or with sudo')
292
295
  end
@@ -265,7 +265,7 @@ module ComputeUnit
265
265
  def self.find_all(use_opencl = false)
266
266
  require 'compute_unit/gpus/amd_gpu'
267
267
  require 'compute_unit/gpus/nvidia_gpu'
268
- g = ComputeUnit::AmdGpu.find_all(use_opencl) + ComputeUnit::NvidiaGpu.find_all(use_opencl)
268
+ g = compute_classes.map { |klass| klass.find_all(use_opencl) }.flatten
269
269
  g.sort_by(&:index)
270
270
  end
271
271
 
@@ -90,6 +90,16 @@ module ComputeUnit
90
90
  utilization
91
91
  end
92
92
 
93
+ # @return [Integer] - the temperature of the asic chip
94
+ def asic_temp
95
+ read_hwmon_data('temp2_input', 0).to_i / 1000
96
+ end
97
+
98
+ # @return [Integer] - temperature of the memory
99
+ def mem_temp
100
+ read_hwmon_data('temp3_input', 0).to_i / 1000
101
+ end
102
+
93
103
  # @return [Integer] - returns temp of gpu in celius
94
104
  def temp
95
105
  read_hwmon_data('temp1_input', 0).to_i / 1000
@@ -421,16 +431,6 @@ module ComputeUnit
421
431
  vddci_voltage_table.first[:volt]
422
432
  end
423
433
 
424
- # @return [Integer] - the temperature of the asic chip
425
- def asic_temp
426
- read_hwmon_data('temp2_input', 0).to_i / 1000
427
- end
428
-
429
- # @return [Integer] - the temperature of the memory chips
430
- def mem_temp
431
- read_hwmon_data('temp3_input', 0).to_i / 1000
432
- end
433
-
434
434
  def set_mem_clock_and_vddc(mem_clock, mem_volt)
435
435
  return unless experimental_on?
436
436
 
@@ -8,6 +8,7 @@ module ComputeUnit
8
8
  MAKE = 'Nvidia'
9
9
  SUBTYPE = 'nvidia'
10
10
  NVIDIA_SMI = '/usr/bin/nvidia-smi'
11
+ NVIDIA_PROC_PATH = ENV['NVIDIA_PROC_PATH'] || File.join(ComputeUnit::Device::PROC_PATH, 'driver', 'nvidia', 'gpus')
11
12
 
12
13
  def initialize(device_path, opts = {})
13
14
  data = self.class.read_information_file(device_path).merge(opts)
@@ -159,7 +160,7 @@ module ComputeUnit
159
160
  def information_file
160
161
  @information_file ||= begin
161
162
  device_name = File.basename(device_path)
162
- File.join('/proc/driver/nvidia/gpus', device_name, 'information')
163
+ File.join(NVIDIA_PROC_PATH, device_name, 'information')
163
164
  end
164
165
  end
165
166
 
@@ -209,7 +210,8 @@ module ComputeUnit
209
210
  # :device_minor=>"7"}
210
211
  def self.read_information_file(device_path)
211
212
  device_name = File.basename(device_path)
212
- information_file = "/proc/driver/nvidia/gpus/#{device_name}/information"
213
+ information_file = File.join(NVIDIA_PROC_PATH, device_name, 'information')
214
+
213
215
  File.open(information_file, 'r') do |file|
214
216
  content = file.read
215
217
  content.scan(/\n?([\w\s]*):\s+(.*)/).map { |key, value| [key.downcase.tr(' ', '_').to_sym, value] }.to_h
@@ -8,7 +8,7 @@ module ComputeUnit
8
8
  if ENV['LOG_FILENAME'] && File.exist?(ENV['LOG_FILENAME'])
9
9
  ENV['LOG_FILENAME']
10
10
  else
11
- STDOUT
11
+ STDERR
12
12
  end
13
13
  end
14
14
 
@@ -18,7 +18,7 @@ module ComputeUnit
18
18
  log.level = log_level
19
19
  log.progname = 'ComputeUnit'
20
20
  log.formatter = proc do |severity, datetime, progname, msg|
21
- if Logger.log_file == STDOUT
21
+ if Logger.log_file == STDERR
22
22
  "#{severity} - #{progname}: #{msg}\n".send(color(severity))
23
23
  else
24
24
  "#{datetime} #{severity} - #{progname}: #{msg}\n".send(color(severity))
@@ -1,5 +1,5 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module ComputeUnit
4
- VERSION = '0.1.0'
4
+ VERSION = '0.3.2'
5
5
  end
metadata CHANGED
@@ -1,14 +1,14 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: compute_unit
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.1.0
4
+ version: 0.3.2
5
5
  platform: ruby
6
6
  authors:
7
7
  - Corey Osman
8
8
  autorequire:
9
9
  bindir: exe
10
10
  cert_chain: []
11
- date: 2020-05-25 00:00:00.000000000 Z
11
+ date: 2020-10-28 00:00:00.000000000 Z
12
12
  dependencies:
13
13
  - !ruby/object:Gem::Dependency
14
14
  name: opencl_ruby_ffi
@@ -16,42 +16,42 @@ dependencies:
16
16
  requirements:
17
17
  - - "~>"
18
18
  - !ruby/object:Gem::Version
19
- version: 1.3.0
19
+ version: 1.3.4
20
20
  type: :runtime
21
21
  prerelease: false
22
22
  version_requirements: !ruby/object:Gem::Requirement
23
23
  requirements:
24
24
  - - "~>"
25
25
  - !ruby/object:Gem::Version
26
- version: 1.3.0
26
+ version: 1.3.4
27
27
  - !ruby/object:Gem::Dependency
28
28
  name: bundler
29
29
  requirement: !ruby/object:Gem::Requirement
30
30
  requirements:
31
31
  - - "~>"
32
32
  - !ruby/object:Gem::Version
33
- version: '2.0'
33
+ version: '2.1'
34
34
  type: :development
35
35
  prerelease: false
36
36
  version_requirements: !ruby/object:Gem::Requirement
37
37
  requirements:
38
38
  - - "~>"
39
39
  - !ruby/object:Gem::Version
40
- version: '2.0'
40
+ version: '2.1'
41
41
  - !ruby/object:Gem::Dependency
42
42
  name: rake
43
43
  requirement: !ruby/object:Gem::Requirement
44
44
  requirements:
45
45
  - - "~>"
46
46
  - !ruby/object:Gem::Version
47
- version: '10.0'
47
+ version: '13'
48
48
  type: :development
49
49
  prerelease: false
50
50
  version_requirements: !ruby/object:Gem::Requirement
51
51
  requirements:
52
52
  - - "~>"
53
53
  - !ruby/object:Gem::Version
54
- version: '10.0'
54
+ version: '13'
55
55
  - !ruby/object:Gem::Dependency
56
56
  name: rspec
57
57
  requirement: !ruby/object:Gem::Requirement
@@ -68,9 +68,9 @@ dependencies:
68
68
  version: '3.0'
69
69
  description: |2+
70
70
 
71
- A ruby library that searches uses the linux sysfs file system for compute unit devices such as
72
- CPUS, GPUs and other ASIC compute devices. Allows programmatic access collect real time metrics from the kernel or relatated driver toolchain.
73
- Is meant to be used as a toolchain to future build tooling on. This library also makes use of opencl toolchain and requires
71
+ A ruby library that searches the linux sysfs file system for compute unit devices such as
72
+ CPUS, GPUs and other ASIC compute devices. Allows programmatic access to collect real time metrics from the kernel or relatated driver toolchain.
73
+ Is meant to be used as a toolchain for future tools. This library also makes use of opencl library and requires
74
74
  the opencl_ruby_ffi gem.
75
75
 
76
76
  email:
@@ -119,7 +119,7 @@ licenses:
119
119
  metadata:
120
120
  homepage_uri: https://gitlab.com/blockops/compute_unit
121
121
  source_code_uri: https://gitlab.com/blockops/compute_unit
122
- changelog_uri: https://gitlab.com/blockops/compute_unit/CHANGELOG
122
+ changelog_uri: https://gitlab.com/blockops/compute_unit/-/blob/master/CHANGELOG.md
123
123
  post_install_message:
124
124
  rdoc_options: []
125
125
  require_paths: