240 bytes of overhead in PXI-5152

[galp]

The manual for the PXI5152 states that for each recorded waveform the amount of memory required is the number of samples + 240 bytes rounded up to the nearest 128.

Does anyone know what is stoed in the 240 bytes?  We are using the card to trigger 100's of thousands of small waveforms (<64 samples) per second and this large amount of overhead is haveing a huge impact on our overall performance.

I am wondering if there is anything that can be done to eliminate this overhead for applications where very small waveforms are needed.

Thanks for any feedback!

Paul

[Dan_K]

Hello Paul,

The 240 bytes for each record is a header file that contains all of the metadata for each record. This cannot be changed. You are going to have to either reduce the number of records you acquire each second or acquire for a smaller amount of time in order to not fill up your onboard memory.

[Dan_K · Accepted solution]

Hello Paul,

I have more information on the exact purpose and nature of the extra bytes per record fetched. First, this extra data is not the record metadata. This data is returned to the host computer asynchronously and is not stored in the board memory with the record.

Next, the actual amount of data stored in the record in excess of the actual samples is actually 400 bytes on the NI 5152 (This information was corrected in the September 2008 version of the NI PXI/PCI-5152 Specifications, page 15). These bytes are allocated to each record to allow for proper operation of the board. This is necessary to allow flushing of the datapath to retrieve all of the data and to protect the record from being overwritten.