Multiarchitecture Programming Becoming the Norm

Santa Cruz, CA, June 24, 2024 –The issue of optimizing for CPUs vs GPUs is becoming moot as over half, 54%, of developers either are currently designing apps to account for processor specifics across multiple hardware architectures including CPUs, GPUs, or accelerators, or they have done so in the past according to Evans Data’s recently released Global Development Survey. This comprehensive survey of software developers worldwide was conducted in six languages across four regions and also showed that of those developers experienced with multiarchitecture 79% use SaaS-based cloud products to support this, and that usage was especially strong in the APAC region.

The survey also probed for what the key benefits are from programming for multiple architectures. The top benefit cited lies in the flexibility to leverage specialized architectures for specific tasks. Developers can exploit the unique strengths of various hardware platforms using multiarchitecture programming. For instance, computationally intensive tasks might be targeted towards execution on GPUs that excel in parallel processing, while other functionalities can utilize the general-purpose capabilities of CPUs. This flexibility was particularly valued in Latin America.

Evans Data Corp’s biannual Global Development survey provides fresh data on current a range of current development topics. Its scope includes AI and machine learning, Cloud development, HPC, Hardware platforms, Block Chain, Operating systems and Language Use, Security, Quantum development, and more. For an in-depth view of the topics addressed, view the full table of contents and more at