⚙ Configuration

Choose the mode and the memory block size to test.

Mode:
Block:
PROGRESS 0%
Tested Memory Map 0 / 64 blocks
Free
Writing
Reading
OK
Error
📦
Allocation
Reserves the memory block for the test
WAITING
✏️
Write
Writes bit patterns across the entire memory
WAITING
🔍
Read and Verify
Reads and checks every byte previously written
WAITING
🔄
Simultaneous Read/Write
Measures bandwidth with mixed operations
WAITING
🛡️
0xAA/0x55 Integrity
Alternating pattern for bit-fault detection
WAITING
Bandwidth (MB/s) 0 MB/s
Test Log
--:--:--Waiting for test to start...
Available RAM (est.)
via performance.memory
Write Speed
MB/s (peak)
Read Speed
MB/s (peak)
Total Bandwidth
write + read
Errors Detected
0
no faults
Block / Mode
32 MB
mode: quick
JavaScript Heap
Used
Total
📋 RAM Test Result
Write (peak)
Read (peak)
Total Bandwidth
Tested Block
Errors
0
Integrity
Memory Score
⚠️

Important notice: This site and all of its tests are provided free of charge and on an experimental basis. We are not responsible for any damage, instability, crashing or overheating caused to your computer during or after running the tests. Use responsibly — if your system shows signs of instability, stop the test immediately.

How it works: Writing uses 4 simultaneous strategiesUint32Array.fill(0xDEADBEEF), fill(0), Float64Array.fill(0) and fill(0xFFFFFFFF) — and reports the fastest. Native fill() is compiled by the JIT into the processor's memset/AVX instructions, measuring real bandwidth. Reading walks every element with Float64Array (8 bytes/access). RW uses f64.set(f64.subarray()), which V8 compiles into a vectorized memmove. A double warm-up primes the JIT before measuring. Close other tabs for more accurate results.