Achronix ACE Version 5.0 User Manual
Page 246
Reports
Chapter 3. Concepts
Timing Results Summary Section
The Timing Results Summary section of the report will only be generated if either the ”Run Post-Route
Timing Analysis” or ”Run Sign-off Timing Analysis”
are enabled in the
when
Multiprocess execution is started.
Achieved Frequency cells are color-coded to indicate whether the Target Frequency was achieved for each
implementation’s defined clock domains. If the Achieved Frequency meets the Target Frequency for all
clocks in an implementation, the Implementation Name cell will also be color coded green to indicate
success. If errors are encountered during flow execution, the appropriate Implementation Name cell will be
colored red, and ”(Flow Error) ” will also appear below the implementation name.
While an implementation is waiting to be executed, or is currently executing, the Clock Domain column
will contain a message to indicate the implementation’s execution status.
Note:
If the ”Run Post-Route Timing Analysis” and ”Run Sign-off Timing Analysis” flow steps
are both enabled, but errors are encountered within an implementation after Post-Route
Timing Analysis is complete but before Sign-off Timing Analysis is complete, the results
included in the table for that implementation will be from Post-Route Timing Analysis,
not from Sign-off Timing Analysis, despite what the column heading indicates.
Synchronous Fabrics
(Includes the 22iHD device family)
If all selected implementations used synchronous fabrics, the extra columns relevant to asynchronous fabrics
will be omitted from the report.
Asynchronous Fabrics
(Includes the 22iHP device family)
For asynchronous target devices, data for the Current Results and Potential Results columns may be
independently enabled and disabled for each implementation as Implementation Options. Within the
, under the ”Timing Analysis” category, they are called Part I: Current Performance and Part
II: Potential Performance
, respectively. (These options are only available for implementations targeting
asynchronous fabrics.)
Runtime Results Summary Section
This section of the report will always be generated, and will indicate the total flow execution time
26
for each
implementation.
Rows in the table will indicate whether an implementation’s flow execution is incomplete (running or
waiting to run) or has encountered errors.
Users should be aware that implementation flow execution times will vary depending upon the available
processors, available memory, and total load on the workstation executing ACE. An implementation’s
selected
(and thus the
) will also have an impact upon implementation
runtimes.
If comparing runtimes on multi-user workstations, please remain aware that workstation load may vary
widely over the multiprocess execution period, making direct runtime comparisons difficult.
Even on single-user workstations, if using a Parallel Queue Count greater than one, be aware that the
last-executed implementation(s) will likely be executing (at least partially) with a reduced machine load
compared to the first-executed implementations. (As the parallel execution queue is exhausted, new
implementation processes are not started, thus fewer processes are executing, meaning that the last-
executed implementations have the lowest contention for processing cores, memory, I/O, etc.)
Because of all of these limitations, the reported implementation flow execution times are meant to be used
only as general guidelines, not as benchmarking times.
26
For those experienced with benchmarking, the reported times are wall-clock times, not cputimes.
UG001 Rev. 5.0 - 5th December 2012
234