Documentation
¶
Overview ¶
Package benchmark provides benchmark management for the CipherSwarm agent. It handles running hashcat benchmarks, submitting results to the server, and caching results for retry on failure.
Index ¶
- type Config
- type Manager
- func (m *Manager) RunBackgroundBenchmarks(ctx context.Context, isIdle func() bool)
- func (m *Manager) RunCapabilityDetection(ctx context.Context) ([]Result, error)
- func (m *Manager) SubmitCapabilityResults(ctx context.Context, results []Result) error
- func (m *Manager) TrySubmitCachedBenchmarks(ctx context.Context) bool
- func (m *Manager) UpdateBenchmarks(ctx context.Context) error
- type Result
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Config ¶ added in v0.6.2
type Config struct {
// OutPath is the directory for hashcat output and charset temp files.
OutPath string
// ZapsPath is the directory for zap files, removed on cleanup unless retained.
ZapsPath string
// RetainZapsOnCompletion keeps the zaps directory after a session completes.
RetainZapsOnCompletion bool
// EnableAdditionalHashTypes enables all hash types in full benchmark mode.
EnableAdditionalHashTypes bool
}
Config holds injected path and benchmark-mode configuration for a Manager. It is a value type (safe to copy), mirroring task.Config and devices.DeviceConfig, so benchmark sessions are constructable without reading agentstate directly.
type Manager ¶
type Manager struct {
DeviceConfig devices.DeviceConfig
// Config holds injected path and benchmark-mode configuration for this Manager.
Config Config
// contains filtered or unexported fields
}
Manager handles benchmark operations including running benchmarks, submitting results, and managing the benchmark cache.
func NewManager ¶
func NewManager(agentsClient api.AgentsClient) *Manager
NewManager creates a new benchmark Manager with the given API client.
func (*Manager) RunBackgroundBenchmarks ¶ added in v0.6.2
RunBackgroundBenchmarks replaces placeholder cache entries with real benchmark results by running hashcat --benchmark -m <type> for each placeholder hash type. It waits for the agent to be idle (per the injected isIdle predicate) before each run and submits results incrementally. The method is designed to run as a long-lived goroutine.
func (*Manager) RunCapabilityDetection ¶ added in v0.6.2
RunCapabilityDetection runs hashcat --hash-info --machine-readable to discover supported hash types without executing a full benchmark. It returns placeholder Result entries (SpeedHs="1", Placeholder=true) for each discovered type.
func (*Manager) SubmitCapabilityResults ¶ added in v0.6.2
SubmitCapabilityResults caches and submits placeholder capability-detection results. It is the deferred-benchmark equivalent of the full UpdateBenchmarks path.
func (*Manager) TrySubmitCachedBenchmarks ¶
TrySubmitCachedBenchmarks attempts to submit previously cached benchmark results to the server. Returns false immediately if the force-benchmark flag is set (stale results should not be submitted). If all cached results are already marked as submitted, returns true immediately without re-persisting the cache. Otherwise, filters to only unsubmitted results, sends them, marks as submitted, and persists the updated cache. Returns true if all results are now submitted, false otherwise (cache is preserved for the next attempt).
func (*Manager) UpdateBenchmarks ¶
UpdateBenchmarks updates the benchmark metrics using Hashcat. It first checks for cached results from a previous run. If a valid cache exists (and the force-benchmark flag is not set), it attempts submission of only unsubmitted results. Submission failure of cached results is non-fatal — it returns nil and the cache is preserved for retry via TrySubmitCachedBenchmarks.
When no cache exists (or force re-run is requested), it runs a new benchmark session and delegates to cacheAndSubmitBenchmarks, which may return an error if both the cache save and submission fail simultaneously.
type Result ¶ added in v0.6.2
type Result struct {
Device string `json:"device,omitempty"` // Device is the numeric device ID used for the benchmark.
DeviceName string `json:"device_name,omitempty"` // DeviceName is the human-readable device name (display-only, not sent to API).
HashType string `json:"hash_type,omitempty"` // HashType is the type of hash used for the benchmark.
RuntimeMs string `json:"runtime,omitempty"` // RuntimeMs is the runtime of the benchmark in milliseconds.
HashTimeMs string `json:"hash_time,omitempty"` // HashTimeMs is the time taken to hash in milliseconds.
SpeedHs string `json:"hash_speed,omitempty"` // SpeedHs is the hash speed in hashes per second.
Submitted bool `json:"submitted,omitempty"` // Submitted indicates whether this result has been accepted by the server.
Placeholder bool `json:"placeholder,omitempty"` // Placeholder indicates this is a capability-detection result, not a real benchmark.
}
Result represents the outcome of a benchmark session.
Source Files