workload_profile
Workload profile
Workload profile: Cache Intensive
Precomputed compound score for Cache Intensive workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 50% RedisRedisAn open-source, in-memory key-value data store, used as a database, cache, and message broker. It keeps its data in RAM, supports data structures such as strings, hashes, lists, and sets, and can persist data to disk with periodic snapshots or an append-only log.Redis docsGlossary RPS (pipeline=1, SET), 20% Redis RPS (pipeline=16, SET), 10% PassMarkPassMarkThe PerformanceTest benchmark suite of PassMark Software, for Windows, Linux, macOS, and mobile devices. It runs fixed CPU, memory, disk, and graphics tests, such as math workloads involving compression, encryption, and physics simulations, and combines the results into an overall PassMark Rating.PerformanceTestGlossary Memory MarkMemory MarkThe overall memory score of PassMark PerformanceTest. It combines the results of the memory tests, such as database operations, cached and uncached reads, writes, and latency, into one unitless number, where higher is better.PassMark Memory BenchmarksGlossary (composite), 10% Memory bandwidth (read, 16 MB ~ L3), 10% PassMark single-thread CPUCPU: Central Processing UnitThe general-purpose processor of a computer, which runs the operating system and programs. A CPU has one or more physical cores, and with SMT each core runs two or more hardware threads.WikipediaGlossary. Rationale for component selection: In-memory key-value store workload, mixing direct Redis performance metrics with memory speed and latency benchmarks, and single-core CPU performance profiles.
Benchmark details
- Benchmark ID
workload_profile:cache- Unit of measurement
- —
- Server types evaluated
- 3,735
- Total measurements
- 3,735
- Range of measurements
- 0.011 – 15
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| Redis RPS (pipeline=1, SET) | redis:rps-extrapolated | pipeline=1 operation=SET | 50% |
| Redis RPS (pipeline=16, SET) | redis:rps-extrapolated | pipeline=16 operation=SET | 20% |
| PassMark Memory Mark (composite) | passmark:memory_mark | — | 10% |
| Memory bandwidth (read, 16 MB ~ L3) | bw_mem | size=16 operation=rd | 10% |
| PassMark single-thread CPU | passmark:cpu_single_threaded_test | — | 10% |
Score distribution (3,735 measurements)
Workload profile: CI/CD Build
Precomputed compound score for CI/CD Build workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 50% GeekbenchGeekbenchA cross-platform CPU and GPU compute benchmark by Primate Labs. Its CPU benchmark runs workloads modeled on real applications, such as file compression, image processing, and web browsing, first on one core and then on all cores, and reports a Single-Core and a Multi-Core score. Geekbench 6 scores are calibrated so that a baseline system scores 2500.GeekbenchGlossary Clang compilation (multi-core), 10% Geekbench Clang compilation (single-core), 20% stress-ngstress-ngAn open-source tool that loads a computer in selectable ways, with hundreds of stress tests (stressors) that exercise the CPU, memory, file systems, and kernel interfaces. With --metrics, it reports each stressor's throughput in bogo ops/s.stress-ng on GitHubGlossary div16 best-N cores, 5% PassMark integer math, 5% PassMark compression, 5% Brotli compression (multi-core, level 0), 5% PassMark string sorting. Rationale for component selection: Build performance is mainly driven by multi-core compilation throughput, but also bundles single-core compilation speed and general CPU performance, multi-core compression and text/scripting processing.
Benchmark details
- Benchmark ID
workload_profile:cicd- Unit of measurement
- —
- Server types evaluated
- 3,742
- Total measurements
- 3,742
- Range of measurements
- 0.0085 – 18
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| Geekbench Clang compilation (multi-core) | geekbench:clang | cores=multi | 50% |
| Geekbench Clang compilation (single-core) | geekbench:clang | cores=single | 10% |
| stress-ng div16 best-N cores | stress_ng:bestn | — | 20% |
| PassMark integer math | passmark:cpu_integer_maths_test | — | 5% |
| PassMark compression | passmark:cpu_compression_test | — | 5% |
| Brotli compression (multi-core, level 0) | compression_text:compress | algo=brotli cores=single compression_level=0 | 5% |
| PassMark string sorting | passmark:cpu_string_sorting_test | — | 5% |
Score distribution (3,742 measurements)
Workload profile: Compute Heavy Applications
Precomputed compound score for Compute Heavy Applications workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 15% stress-ng div16 best-N cores, 10% stress-ng div16 single core, 20% PassMark CPU MarkCPU MarkThe overall CPU score of PassMark PerformanceTest. It combines the results of the CPU tests, such as integer and floating-point math, prime numbers, sorting, encryption, compression, physics, and a single-threaded test, into one unitless number, where higher is better. PassMark Software publishes CPU Mark results submitted by users for thousands of processors.PassMark CPU BenchmarksGlossary (composite), 10% Memory bandwidth (read, 64 MB), 15% PassMark floating point, 15% PassMark AVX/SSE/FMA (SIMD), 10% PassMark integer math, 5% PassMark physics simulation. Rationale for component selection: Number-crunching workload augmenting raw CPU performance stressing, general CPU performance benchmarks, memory bandwidth, and pure math computation speed like floating point, integer, SIMD (AVX/SSE/FMA) operations.
Benchmark details
- Benchmark ID
workload_profile:compute- Unit of measurement
- —
- Server types evaluated
- 3,742
- Total measurements
- 3,742
- Range of measurements
- 0.011 – 16
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| stress-ng div16 best-N cores | stress_ng:bestn | — | 15% |
| stress-ng div16 single core | stress_ng:best1 | — | 10% |
| PassMark CPU Mark (composite) | passmark:cpu_mark | — | 20% |
| Memory bandwidth (read, 64 MB) | bw_mem | size=64 operation=rd | 10% |
| PassMark floating point | passmark:cpu_floating_point_maths_test | — | 15% |
| PassMark AVX/SSE/FMA (SIMD) | passmark:cpu_extended_instructions_test | — | 15% |
| PassMark integer math | passmark:cpu_integer_maths_test | — | 10% |
| PassMark physics simulation | passmark:cpu_physics_test | — | 5% |
Score distribution (3,742 measurements)
Workload profile: Data Analysis
Precomputed compound score for Data Analysis workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 70% PassMark CPU MarkCPU MarkThe overall CPU score of PassMark PerformanceTest. It combines the results of the CPU tests, such as integer and floating-point math, prime numbers, sorting, encryption, compression, physics, and a single-threaded test, into one unitless number, where higher is better. PassMark Software publishes CPU Mark results submitted by users for thousands of processors.PassMark CPU BenchmarksGlossary (composite), 10% Gzip compression (single-core, level 5), 10% Memory bandwidth (read, 64 MB), 10% PassMark Memory MarkMemory MarkThe overall memory score of PassMark PerformanceTest. It combines the results of the memory tests, such as database operations, cached and uncached reads, writes, and latency, into one unitless number, where higher is better.PassMark Memory BenchmarksGlossary (composite). Rationale for component selection: Data analysis and ETL workloads are memory-bandwidth-bound and CPU-throughput-driven. The profile combines general CPU performance and memory bandwidth/latency as the primary drivers, supplemented by single-core compression speed as a proxy for serialisation-heavy ETL tasks.
Benchmark details
- Benchmark ID
workload_profile:data_analysis- Unit of measurement
- —
- Server types evaluated
- 3,711
- Total measurements
- 3,711
- Range of measurements
- 0.013 – 6.3
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| PassMark CPU Mark (composite) | passmark:cpu_mark | — | 70% |
| Gzip compression (single-core, level 5) | compression_text:compress | algo=gzip cores=single compression_level=5 | 10% |
| Memory bandwidth (read, 64 MB) | bw_mem | size=64 operation=rd | 10% |
| PassMark Memory Mark (composite) | passmark:memory_mark | — | 10% |
Score distribution (3,711 measurements)
Workload profile: Relational Database inactive
Precomputed compound score for Relational Database workloads. A weighted average of normalised scores, each normalised to [0, 1] across all servers in the dataset. Component weights: 35% PassMark in-memory DB operations, 30% PassMark memory latency (512 MB), 20% PassMark Memory MarkMemory MarkThe overall memory score of PassMark PerformanceTest. It combines the results of the memory tests, such as database operations, cached and uncached reads, writes, and latency, into one unitless number, where higher is better.PassMark Memory BenchmarksGlossary (composite), 15% PassMark single-thread CPU. Rationale for component selection: Relational database workload (PostgreSQLPostgreSQLAn open-source object-relational database system, developed by a global community and descended from the POSTGRES project at UC Berkeley in 1986. It uses and extends SQL, with transactions, many index types, extensions, and JSON data.PostgreSQL docsGlossary, MySQL, transactional OLTP). Direct DB operation throughput is the primary driver, followed by memory latency for index lookups and buffer pool access, memory subsystem performance for working-set throughput, and single-thread CPU for query execution.
Benchmark details
- Benchmark ID
workload_profile:database- Unit of measurement
- —
- Server types evaluated
- —
- Total measurements
- —
- Range of measurements
- —
- Better
- ↑ Higher is better
Workload profile: LLM Inference
Precomputed compound score for LLMLLM: Large Language ModelA neural network, typically with billions of parameters, trained on large amounts of text to predict the next token. It can generate, summarize and translate text, answer questions and write code.WikipediaGlossary Inference workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 15% LLM text generationtext generationThe second phase of LLM inference, also called decoding, in which the model produces output tokens one at a time, each depending on the tokens before it. Every step reads all model weights and the key-value cache, so this phase is usually limited by memory bandwidth rather than compute, and runs much slower per token than prompt processing.Glossary (SmolLM-135M, 128 tok), 15% LLM prompt processingprompt processingThe first phase of LLM inference, also called prefill, in which the model reads all tokens of the input prompt and stores their intermediate results in its key-value cache. The prompt's tokens are processed together in batches, so this phase is usually limited by compute and runs much faster per token than text generation. Its speed decides how long a long input takes before the first output token appears.Glossary (SmolLM-135M, 512 tok), 15% LLM text generation (Llama 7B, 128 tok), 15% LLM prompt processing (Llama 7B, 512 tok), 15% LLM text generation (Llama-3.3 70B, 128 tok), 15% LLM prompt processing (Llama-3.3 70B, 512 tok), 5% Memory bandwidth (read, 256 MB), 2% PassMark AVX/SSE/FMA (SIMD), 2% PassMark floating point. Rationale for component selection: VRAM and memory-bandwidth-bound LLM inference workload, using direct LLM speed benchmarks at three model sizes, and supplementing with raw memory bandwidth and SIMD performance benchmarks.
Benchmark details
- Benchmark ID
workload_profile:llm- Unit of measurement
- —
- Server types evaluated
- 3,416
- Total measurements
- 3,416
- Range of measurements
- 0.0045 – 19
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| LLM text generation (SmolLM-135M, 128 tok) | llm_speed:text_generation | model=SmolLM-135M.Q4_K_M.gguf tokens=128 | 15% |
| LLM prompt processing (SmolLM-135M, 512 tok) | llm_speed:prompt_processing | model=SmolLM-135M.Q4_K_M.gguf tokens=512 | 15% |
| LLM text generation (Llama 7B, 128 tok) | llm_speed:text_generation | model=llama-7b.Q4_K_M.gguf tokens=128 | 15% |
| LLM prompt processing (Llama 7B, 512 tok) | llm_speed:prompt_processing | model=llama-7b.Q4_K_M.gguf tokens=512 | 15% |
| LLM text generation (Llama-3.3 70B, 128 tok) | llm_speed:text_generation | model=Llama-3.3-70B-Instruct-Q4_K_M.gguf tokens=128 | 15% |
| LLM prompt processing (Llama-3.3 70B, 512 tok) | llm_speed:prompt_processing | model=Llama-3.3-70B-Instruct-Q4_K_M.gguf tokens=512 | 15% |
| Memory bandwidth (read, 256 MB) | bw_mem | size=256 operation=rd | 5% |
| PassMark AVX/SSE/FMA (SIMD) | passmark:cpu_extended_instructions_test | — | 3% |
| PassMark floating point | passmark:cpu_floating_point_maths_test | — | 3% |
Score distribution (3,416 measurements)
Workload profile: ML inference inactive
Precomputed compound score for CPU-based machine-learning inference workloads. A weighted average of normalised scores, each normalised to [0, 1] across all servers in the dataset. Component weights: 20% LLM text generation (llama-7b, 128 tokenstokenThe unit of text that an LLM reads and writes, as defined by the model's tokenizer, such as a word, part of a word, or a punctuation mark. Prompt and output lengths are counted in tokens, and LLM inference speed is usually measured in tokens per second.Glossary), 15% memory bandwidth read (256 MB, bw_membw_memThe memory bandwidth tool of LMbench. It moves data through a buffer of a given size in one of several modes, such as rd (read and sum an array), wr (write a constant to each element), rdwr (read, then write the same location), and the copy modes cp, fcp, and bcopy. It reports the buffer size in MB and the bandwidth in MB/s; with -P, it runs several processes, each with its own buffer, and reports their total bandwidth.`bw_mem` manualGlossary), 10% LLM prompt processing (llama-7b, 512 tokens), 10% LLM text generation (gemma-2b, 128 tokens), 10% PassMark Memory MarkMemory MarkThe overall memory score of PassMark PerformanceTest. It combines the results of the memory tests, such as database operations, cached and uncached reads, writes, and latency, into one unitless number, where higher is better.PassMark Memory BenchmarksGlossary (composite), 10% PassMark AVX/SSE/FMA (SIMD), 10% Geekbench object detection (multi-core), 5% PassMark floating point, 5% Geekbench background blur (multi-core), 5% Geekbench structure-from-motion (multi-core).
Benchmark details
- Benchmark ID
workload_profile:ml- Unit of measurement
- —
- Server types evaluated
- —
- Total measurements
- —
- Range of measurements
- —
- Better
- ↑ Higher is better
Workload profile: Web Server
Precomputed compound score for Web Server workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 30% Static web RPS (1 KiB, 8 conn/vCPU), 20% Static web RPS (64 KiB, 8 conn/vCPU), 20% Static web throughput (256 KiB, 8 conn/vCPU), 20% OpenSSL AES-256-CBC (16 kB blocks), 5% Gzip compression (multi-core, level 5), 5% PassMark string sorting. Rationale for component selection: Primary workloads drivers are single-process static HTTPHTTP: Hypertext Transfer ProtocolThe request-response protocol of the web, in which a client asks a server for a resource by URL and gets back a status code, headers, and a body. HTTPS is HTTP over an encrypted connection.MDNGlossary serving speed and throughput, text processing, TLS termination, and asset compression.
Benchmark details
- Benchmark ID
workload_profile:web- Unit of measurement
- —
- Server types evaluated
- 3,738
- Total measurements
- 3,738
- Range of measurements
- 0.011 – 13
- Better
- ↑ Higher is better
Components
| Component | Benchmark | Configuration | Weight |
|---|---|---|---|
| Static web RPS (1 KiB, 8 conn/vCPU) | static_web:rps-extrapolated | size=1k connections_per_vcpus=8 | 30% |
| Static web RPS (64 KiB, 8 conn/vCPU) | static_web:rps-extrapolated | size=64k connections_per_vcpus=8 | 20% |
| Static web throughput (256 KiB, 8 conn/vCPU) | static_web:throughput-extrapolated | size=256k connections_per_vcpus=8 | 20% |
| OpenSSL AES-256-CBC (16 kB blocks) | openssl | algo=AES-256-CBC block_size=16384 | 20% |
| Gzip compression (multi-core, level 5) | compression_text:compress | algo=gzip cores=multi compression_level=5 | 5% |
| PassMark string sorting | passmark:cpu_string_sorting_test | — | 5% |
Score distribution (3,738 measurements)