Lossless data compression is highly desirable in enterprise and cloud environments for storage and memory cost savings and improved utilization I/O and network. While the value provided by compression is recognized, its application in practice is often limited because it's a processor intensive operation resulting low throughput and high elapsed time for compression intense workloads.The IBM POWER9 and IBM z15 systems overcome the shortcomings of existing approaches by including a novel on-chip integrated data compression accelerator. The accelerator reduces processor cycles, I/O traffic, memory and storage footprint of many applications practically with zero hardware cost. The accelerator also eliminates the cost and I/O slots that would have been necessary with FPGA/ASIC based compression adapters. On the POWER9 chip, a single accelerator uses less than 0.5% of the processor chip area, but provides a 388x speedup factor over the zlib compression software running on a general-purpose core and provides a 13x speedup factor over the entire chip of cores. On a POWER9 system, the accelerators provide an end-to-end 23% speedup to Apache Spark TPC-DS workload compared to the software baseline. The z15 chip doubles the compression rate of POWER9 resulting in even much higher speedup factors over the compression software running on general-purpose cores. On a maximally configured z15 system topology, on-chip compression accelerators provide up to 280 GB/s data compression rate, the highest in the industry. Overall, the on-chip accelerators significantly advance the state of the art in terms of area, throughput, latency, compression ratio, reduced processor utilization, power/energy efficiency, and integration into the system stack.This paper describes the architecture, and novel elements of the POWER9 and z15 compression/decompression accelerators with emphasis on trade-offs that made the on-chip implementation possible.

The IBM z15 processor chip contains a new hardware component to perform DEFLATE compliant compression and decompression. The Integrated Accelerator for zEnterprise Data Compression is based on a high-frequency DEFLATE pipeline and includes a hardware generator for dynamic Huffman tables. Accessible as an architected instruction, this engine has been designed for straightforward exploitation by software and is easily available to any application in the problem state. A brand-new hardware/firmware integration model has been developed to provide this complex functionality without imposing restrictions on data patterns or data sizes and without impacting system responsiveness. This article describes the concept, implementation, and verification of DEFLATE, compliant compression acceleration in z15 across both hardware and firmware. It illustrates various challenges that result from incorporating complex data-dependent and data-intense functionality like DEFLATE as an architected instruction and discusses how solutions in hardware/firmware codesign have been applied to overcome these challenges.

IBM z15 replaces the former I/O attached accelerator for DEFLATE, zEnterprise Data Compression (zEDC) Express, with an on-chip accelerator that can be synchronously accessed via an instruction. The integration of this new accelerator in the z/OS software stack has been designed to maintain a consistent user experience for software packages that used the previous technology, while still allowing the enhanced aspects of the new technology to deliver the additional value. Two different access paths for DEFLATE have been created in z/OS to accomplish both goals. For user space programs that utilize the zlib API, z/OS directly executes the instruction synchronously, which avoids overhead and reduces latency. Authorized users continue to utilize existing infrastructure and have the Service Assist Processors (SAP) perform compression in an asynchronous fashion on their behalf. The SAP receives information about the requested task via a thin and efficient communication path to z/OS, invokes the instruction in a well-defined fashion, and returns the result to z/OS.This article describes the integration of DEFLATE acceleration in z15 into the z/OS software stack in both synchronous and asynchronous mode and presents the resulting performance for selected workloads.