Skip to content

Commit e6623aa

Browse files
committed
review memory subsystem with ASCT
1 parent d58f3c3 commit e6623aa

2 files changed

Lines changed: 22 additions & 14 deletions

File tree

content/learning-paths/servers-and-cloud-computing/memory-subsystem/comparative-analysis.md

Lines changed: 2 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -14,15 +14,7 @@ You now have a complete set of memory subsystem measurements for each of your te
1414

1515
Throughout this Learning Path, you ran individual ASCT benchmarks. To characterize a new system from scratch, run all memory benchmarks at once.
1616

17-
First, set a descriptive hostname for your system.
18-
19-
```bash
20-
sudo hostnamectl set-hostname graviton4-c8g
21-
```
22-
23-
Set a similar hostname on the Graviton2 system, for example `graviton2-c6g`.
24-
25-
Next, run all tests and save the results.
17+
Run all tests and save the results.
2618

2719
```bash
2820
sudo asct run memory loaded-latency --output-dir results_$(hostname)
@@ -56,7 +48,7 @@ Save the comparison as CSV for further analysis:
5648
asct diff results_graviton2-c6g/ results_graviton4-c8g/ --format=csv --output-dir comparison/ --benchmarks ^system-info
5749
```
5850

59-
ASCT writes `diff.csv` to the output directory. It also generates comparison plots for supported benchmarks (latency sweep, bandwidth sweep, loaded latency) as PNG files in the same directory.
51+
ASCT writes `diff.csv` to the output directory.
6052

6153
### Filter the comparison
6254

content/learning-paths/servers-and-cloud-computing/memory-subsystem/system-overview.md

Lines changed: 20 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -12,7 +12,7 @@ Before you can characterize a memory subsystem, you need to understand how the s
1212

1313
This Learning Path uses two example AWS Graviton instances:
1414

15-
- AWS Graviton2, `c6g.metal`, instance with Arm Neoverse N1 cores.
15+
- AWS Graviton2, `c6g.16xlarge`, instance with Arm Neoverse N1 cores.
1616
- AWS Graviton4, `c8g.16xlarge`, instance with Arm Neoverse V2 cores.
1717

1818
Both systems run Ubuntu 24.04 with 64 CPUs, 128 GB RAM, and 32 GB of storage.
@@ -21,7 +21,23 @@ These are just examples, and you can follow the same steps on any Arm Linux syst
2121

2222
## Collect basic system information
2323

24-
Start by recording the kernel version and operating system. Run the following command on your system:
24+
Start by setting a descriptive hostname on each system. This makes it easier to identify results later when you use `$(hostname)` in output directory names.
25+
26+
On the Graviton4 instance, run:
27+
28+
```bash
29+
sudo hostnamectl set-hostname graviton4-c8g
30+
```
31+
32+
On the Graviton2 instance, run:
33+
34+
```bash
35+
sudo hostnamectl set-hostname graviton2-c6g
36+
```
37+
38+
Log out and log back in for the hostname to appear in your shell prompt.
39+
40+
Record the kernel version and operating system. Run the following command on your system:
2541

2642
```bash
2743
uname -a && cat /etc/os-release
@@ -122,8 +138,8 @@ free -h
122138
Both systems have 128 GB RAM with similar output:
123139

124140
```output
125-
. total used free shared buff/cache available
126-
Mem: 125Gi 1.8Gi 124Gi 2.4Mi 686Mi 123Gi
141+
total used free shared buff/cache available
142+
Mem: 123Gi 1.6Gi 121Gi 1.2Mi 1.2Gi 121Gi
127143
```
128144

129145
Graviton2 instances use DDR4 memory. Graviton4 instances use DDR5 memory, which provides higher bandwidth per channel and lower access latency.

0 commit comments

Comments
 (0)