ASPP/2024-heraklion-comp-arch
9
0
Fork 17
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity

define what PCI stands for

Browse source
This commit is contained in:
Tiziano Zito 2024-08-16 19:04:18 +02:00
parent ae8dfa153d
commit a03e56b4a0
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
architecture/README.md
View file

@ -73,7 +73,7 @@
- latency: usually much higher than CPU's ➔ the more parallel threads are run the less the price of high latency is paid (latency "hiding") - latency: usually much higher than CPU's ➔ the more parallel threads are run the less the price of high latency is paid (latency "hiding")
- spatial locality is extremely critical - spatial locality is extremely critical
- A portion of the GPU-RAM is accessible to the CPU ➔ the GPU performs the copies - A portion of the GPU-RAM is accessible to the CPU ➔ the GPU performs the copies
- The PCI-Bus bottleneck: data needs to flow from main (CPU) memory to GPU memory and back! - The PCI-Bus (Peripheral Component Interconnect bus) is the bottleneck: data needs to flow from main (CPU) memory to GPU memory and back!
- Problems on a cluster: the GPU does not really support simultanous multiple users payloads! - Problems on a cluster: the GPU does not really support simultanous multiple users payloads!
# Computer Architecture (a concrete example) # Computer Architecture (a concrete example)
@ -94,7 +94,7 @@ My Laptop:
- Internal clock 650 MHz, 1.54 ns - Internal clock 650 MHz, 1.54 ns
- CAS Latency 34 cycles, Total latency = CAS latency x cycle = 13.09 ns, Throughput 40.6 GB/s - CAS Latency 34 cycles, Total latency = CAS latency x cycle = 13.09 ns, Throughput 40.6 GB/s
- DMI (Direct Media Interface): 8×16 GT/s (≈128 GB/s) - DMI (Direct Media Interface): 8×16 GT/s (≈128 GB/s)
- PCI Express bridges: - PCI (Peripheral Component Interconnect) Express bridges:
- Graphics: 16 GT/s (≈ 8 GB/s) - Graphics: 16 GT/s (≈ 8 GB/s)
- 2× Thunderbolt: 2.5 GT/s (≈ 1 GB/s) and 16 GT/s (≈ 8 GB/s) - 2× Thunderbolt: 2.5 GT/s (≈ 1 GB/s) and 16 GT/s (≈ 8 GB/s)
- GPU Intel Iris, Internal clock 300 Mhz-1.30 GHz, memory 4 GB/2.1 GHz with a bandwidth of 68 GB/s - GPU Intel Iris, Internal clock 300 Mhz-1.30 GHz, memory 4 GB/2.1 GHz with a bandwidth of 68 GB/s