Skip to main content

Optimizations

Eugo applies automatic optimizations to your code to make it run faster and more efficiently, without lifting a finger.

Auto Offload to GPUA Graphics Processing Unit (GPU) is a specialized electronic circuit designed for large-scale parallel processing of numerical, scientific, and graphical data.​

When possible, Eugo will automatically offload your code to the GPUA Graphics Processing Unit (GPU) is a specialized electronic circuit designed for large-scale parallel processing of numerical, scientific, and graphical data. for faster execution. This is especially useful for computationally intensive tasks like matrix multiplication, deep learning, and other parallelizable tasks.

Auto VectorizationVectorization is the process of converting scalar operations into vector operations, allowing multiple data points to be processed simultaneously.​

Eugo will automatically vectorizeVectorization is the process of converting scalar operations into vector operations, allowing multiple data points to be processed simultaneously. your code to take advantage of SIMD instructions on modern CPUs. This can result in significant speedups for certain types of operations, like element-wise arithmetic on large arrays.

Auto Parallelization​

Using Ray, Eugo will automatically parallelize your code to take advantage of multiple CPU cores. This can result in orders of magnitude faster execution times for tasks that can be split into independent subtasks, like map-reduce operations.

Auto JITA Just-In-Time (JIT) compiler is a type of compiler that compiles code at runtime, rather than ahead of time, into machine code to improve performance. Compilation (Python 3.13+)​

Eugo applies the Python 3.13 Just-In-Time (JIT)A Just-In-Time (JIT) compiler is a type of compiler that compiles code at runtime, rather than ahead of time, into machine code to improve performance. compiler to compile your code into machine code for faster execution. This can result in significant speedups, especially for tight loops and numerical computations. Similar to Numba, the Python JITA Just-In-Time (JIT) compiler is a type of compiler that compiles code at runtime, rather than ahead of time, into machine code to improve performance. automatically compiles Python code into native code. On Eugo, however, we compile your code on one node and distribute it across the cluster.