Microbee General Reference

BeeMark

BeeMark is a software utility used for measuring Microbee performance. A total of 8 tests are currently performed by BeeMark v1.0, with each test returning a separate number or "score" that can be used to compare the relative performance of different Microbee models and/or the effect of hardware modifications. The 8 tests are:

Execution speed of some code residing in normal Program RAM
Execution speed of some code residing in PCG RAM
Read speed from normal Program RAM
Write speed to normal Program RAM
Read speed from Screen RAM
Write speed to Screen RAM
Read speed from PCG RAM
Write speed to PCG RAM

The same test code is executed for tests 1 & 2 so these results are directly comparable with each other. Different test code is used for tests 3 to 8 so while these results are also directly comparable with each other, they are not meaningfully comparable with tests 1 & 2.

BeeMark v1.0

BeeMark v1.0 Test Results

These are the test results collected from various Microbee types.

Microbee Type	CPU Clock MHz	Program RAM Exec	PCG RAM Exec	Program RAM Read	Program RAM Write	Screen RAM Read	Screen RAM Write	PCG RAM Read	PCG RAM Write	Notes
Standard ROM based	2	1143	1143	572	572	572	572	572	572	An early ROM based Microbee with a Standard type 1248 Main board and 2MHz CPU clock.
Standard ROM or Disk based	3.375	1931	1931	966	966	966	966	966	966	A later ROM or Disk based Microbee, still fitted with a Standard type 1248 Main board but now with the more common 3.375MHz CPU clock. This type shows an approximate 69% increase in overall performance compared with the earlier 2MHz models, in line with the 69% increase in CPU clock speed.
Standard ROM or Disk based with Colour option	3.375	1671	849, 818	835	835	460, 409	460, 409	460, 409	460, 409	A Standard 3.375MHz machine with the Standard Colour option fitted (ie, board number 1217-xx). Adding the colour option has some significant effects on performance: General processing performance is reduced by approximately 13%. The cause of this reduction is not yet clear. Video processing performance is reduced by 52-58%. This is mainly caused by the introduction of CPU wait states required for the screen deglitch feature. Oddly, two different machines with the same configuration have produced slightly different video performance results. The reason for this is not yet clear, but may be related to the setting of RV1/2 on the colour board.
Standard ROM or Disk based with Screen Deglitch	3.375	1931	431	966	966	293	293	293	293	A Standard 3.375MHz machine fitted with the video deglitch circuit that's integrated into type 1248-6 Main boards. This combination shows: No reduction in general processing performance. A large 70-78% reduction in video processing performance, highlighting that installation of an optional deglitch disable/bypass switch is probably advisable when this circuit is installed. Similar performance characteristics can be expected from the type 8341-2 Screen Deglitch Add-on board and the 50 way Expansion kit, both of which have a very similar deglitch design.
Premium	3.375	1931	1252	966	966	805	787	805	787	The performance reduction associated with Premium model video deglitch is significantly lower than for Standard models.
256TC	3.375	1931	1252	966	966	805	786	805	786	Essentially the same performance characteristics as the Premium
Premium Plus+	3.375	1931	1252	966	966	805	786	805	786	Essentially the same performance characteristics as the Premium
ubee512 --model=p512k	3.375	1931	1931	966	966	966	966	966	966	ubee512 accurately reproduces general processing performance but does not currently emulate CPU wait states that are associated with Standard or Premium video deglitch.
Suzy's 128K Premium CF Core board	12	6875	2208	3438	3438	2254	2254	2254	2254	The fastest known Microbee - 12MHz!