The TAR.GZ accelerator provides hardware-accelerated TAR and/or GZIP operation up to 25X faster than CPU equivalent.
Generated TAR archives are POSIX compliant.
Generated GZIP archives are compliant to the RFC 1952 GZIP File Format Specification.
Using the TAR.GZ Accelerator Function in 3 easy steps
Per-use monetization and specific to the chosen operation
Reduced compression latency
Support high speed compression functions in real time
Simplify high speed compression software architecture with single FPGA instance vs. multiple CPU instances
Simplify compression tasks with single FPGA instance vs. multiple CPU instances
Reduced cost by
significantly decreasing processing time
using a single FPGA instance instead of 10s of CPU instances
paying only for what you consume with cost based on data volume processed
saving operating costs for on-premise deployment by reducing power consumption
Use cases / Applications
Genomics: compress Genome sequence for storage by up to 4x
Networking: reduce network traffic with compressed files
Performance / Benchmark
Our performance numbers are obtained when compressing 4GB text files
AWS EC2 F1
Intel DCP 1.0
CPU (gzip -3)
CPU (gzip -4)
Our performance numbers are measured as end-to-endperformance running locally on the FPGA instance.
The time that elapses between the Start and the Stop corresponds to the FPGA elapsed time information that is returned when running the accelerator executable on the FPGA instance. This information is available after running the accelerator executable.
When using python REST APIs to run the accelerator remotely on FPGA instances, the end-to-end performance may be lower depending on the network between local machine and the remote FPGA instance
The GZIP operation does not support password-protected archive generation
The TAR operation does not support POSIX 1003.1-1990 extra fields
GZIP Operation : 1 coin per 2GB processed
TAR Operation : 1 coin per 16GB processed
TAR.GZIP Operation : the sum of the 2 when combined (equivalent to 9 coins per 16GB processed
Accelize offers 10,000 coins for free (1,000 coins = $1) to use within the first month (no Credit Card required), and 1,000 coins every subsequent month
By default, in Public Cloud infrastructures, Accelize accelerator functions run on on-demand FPGA instances, which usually cost around $1.5/h. Accelerator functions can also run on spot FPGA instances that cost around $0.5/h. Please check the cost of FPGA instances with your CSP
For more details about pricing, please visit our FAQ page