Workload tests identification

Results comparison of HiperSockets versus OSA provides comparison graphs showing performance results (see Figure 1 through Figure 4). Each graph will contain results (Y-axis) for the nine different user count tests (X-axis) used for each workload type (listed at the bottom of the graph).

The MTU size comparison (Figure 1) shows a reduced set of concurrent users, but, unlike the results comparisons, shows all five workload types on a single graph. The X-axis then contains the full name of each test to allow a clear distinction between the different workload types.

The name for each of the tests is in the following forms:

For transactional tests:
category1c-requestsizexresponsesize--users

Where:

category
is RR for the transactional Request & Response tests
1c
stands for using just 1 uperf client for each uperf server
requestsize
is the number of bytes send by the client to the server
responsesize
is the number of bytes returned in the response by the server to the client
users
is the total number of concurrent users (or processes) generating the overall load

For example: rr1c-200x1000--10 describes a Request & Response test sending a 200-byte request and receiving a 1000-byte response, being generated by each of the 10 concurrent users.

For streaming tests:
category-read|writexpayloadsize--users

Where:

category
is STR for the Streaming tests
read|write
denotes the direction in which data flows relative to the client
payloadsize
is the number of bytes in each datagram
users
is the total number of concurrent users (or processes) generating the overall load

For example: str-readx30k--50 describes a streaming test read of 30KB datagrams, being generated by each of the 50 concurrent users.