Difference: SpeedImprovement (1 vs. 5)

Revision 5
22 Sep 2006 - Main.RadekTrebacz
Line: 1 to 1
 
META TOPICPARENT name="TDCReadoutBoard"
Current speed: 1000 Event/sec -> goal: 5000 Event/sec
Line: 9 to 9
 
  • access with copy data to the local variable - 13.464673 MB/s
  • access with copy data to the structure variable(un_long*)(8 var. in the str.) - 0.497151 MB/s
  • access with copy data to the local variable and use var. from the structure - 10.205232 MB/s
Deleted:
<
<
  • access with copy data to the structure variable and use var. from the structure(int) - 0.491132 MB/s
  • NO access to memory, use the structure variable - 0.389481 MB/s
    Conclusions:
  • we had to use the structure via pointers - 0.497151 MB/s -> 16.354994 MB/s
 

Readout proccess - sensitive points

  • read the memory & put data into the application-memory,
Revision 4
30 Jun 2006 - Main.MichaelTraxler
Line: 1 to 1
 
META TOPICPARENT name="TDCReadoutBoard"
Current speed: 1000 Event/sec -> goal: 5000 Event/sec
Line: 19 to 19
 
  • from the application-memory into the shared memory,
  • from the shared memory into the net.
Changed:
<
<

Measurments of time LVL1_busy (in FPGA)

When there is:
  • 11 words per event then we could achieve 80 kHz with LVL1 trigger,
>
>

Measurements of time LVL1_busy (in FPGA)

When there are:
  • 11 words per event, then we can achieve 80 kHz with LVL1 trigger,
 
  • 240 words per event then 10kHz (almost every channel fires two times).
Changed:
<
<
There are possible improvments in state machines in FPGA to achive higher speed.
>
>
There are possible improvements in state machines in FPGA to achieve higher speed.
  -- MarekPalka - 30 Jun 2006
Revision 3
30 Jun 2006 - Main.MarekPalka
Line: 1 to 1
 
META TOPICPARENT name="TDCReadoutBoard"
Current speed: 1000 Event/sec -> goal: 5000 Event/sec
Line: 19 to 19
 
  • from the application-memory into the shared memory,
  • from the shared memory into the net.
Changed:
<
<

Measurments of time LVL1_busy

>
>

Measurments of time LVL1_busy (in FPGA)

  When there is:
  • 11 words per event then we could achieve 80 kHz with LVL1 trigger,
  • 240 words per event then 10kHz (almost every channel fires two times).
Changed:
<
<
There are possible improvments in state machines in FPGA to achive higher speed.

>
>
There are possible improvments in state machines in FPGA to achive higher speed.
-- MarekPalka - 30 Jun 2006
 

-- RadekTrebacz - 19 Dec 2005
Revision 2
29 Jun 2006 - Main.MarekPalka
Line: 1 to 1
 
META TOPICPARENT name="TDCReadoutBoard"
Current speed: 1000 Event/sec -> goal: 5000 Event/sec
Line: 19 to 19
 
  • from the application-memory into the shared memory,
  • from the shared memory into the net.

Added:
>
>

Measurments of time LVL1_busy

When there is:
  • 11 words per event then we could achieve 80 kHz with LVL1 trigger,
  • 240 words per event then 10kHz (almost every channel fires two times).
    There are possible improvments in state machines in FPGA to achive higher speed.

 

-- RadekTrebacz - 19 Dec 2005
 
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Hades Wiki? Send feedback
Imprint (in German)
Privacy Policy (in German)