Search the Community
Showing results for tags 'randomization'.
Found 4 results
-
Hi All, I have a typical use case, where the sequence item w.r.t driver is always the same, but it's view needs to change while writing constraints. The fields of sequence items to write constraints change w.r.t the header type, there are around 25-30 types of headers. What is the best way to create the base sequence item in this case in line with UVM methodology? Example:- bit [32] header; //Actual header as seen by driver The format of the header is however dynamic:- format-1 : bit[31:21] header_type; bit[20:10] field_2_type_1; bit[9:0] field_3_type_1; format-2: bit[31:21] header_type; bit[20:15] field_2_type_2; bit[14:8] field_3_type_2; bit[7:0] field_4_type_2; format-3: bit[31:21] header_type; bit[20:12] field_2_type_3; bit[11:6] field_3_type_3; bit[5:0] field_4_type_3; ....... ...... I currently have though of a packed union of different structures to be one of the solutions for this type of scenario. This union will be a member of the sequence_item class and the structure will be representation of the different formats of header. Is there a better means of creating polymorphic sequence items in UVM. I am ok to try some system verilog hacks in uvm too for a solution. -
Hi All, Just wanted to check if the UVM sequence macros (UVM_DO_ON_WITH and UVM_DO_WITH) which support randomizing the sequence object are yet supported by UVM-SC or not? If not, what is the best way to add some randomization in the framework? I saw a post where there are some snippets (link given below) using scvx_rand_object class but I am not sure where to find this(didn't find it in latest SCV and UVM-SC libraries). http://nascug.org/events/20th/1-NASCUG20-UVMforSystemC-Karsten.pdf Also, I get errors while using plain UVM_DO_ON macros. Is it expected? The same works fine with ahb_wr_seq->start(top_env->agent->ahb_sequencer_inst); I have registered the sequence class with factory already. UVM_DO_ON(ahb_wr_seq,top_env->agent->ahb_sequencer_inst); Error message says: /nfs/iind/disks/ba_bdl_xplore_disk001/am5/UVM_SC/uvm-systemc-1.0-alpha1/src/uvmsc/macros/uvm_sequence_defines.h:150:31: error: expected type-specifier before ‘typeof’ SEQ_OR_ITEM = dynamic_cast< typeof(SEQ_OR_ITEM) >(create_item(objw__, SEQR, "SEQ_OR_ITEM")); ^ /nfs/iind/disks/ba_bdl_xplore_disk001/am5/UVM_SC/uvm-systemc-1.0-alpha1/src/uvmsc/macros/uvm_sequence_defines.h:115:3: note: in expansion of macro ‘UVM_CREATE_ON’ UVM_CREATE_ON(SEQ_OR_ITEM, SEQR) \ ^ /nfs/iind/disks/ba_bdl_xplore_disk001/am5/UVM_SC/uvm-systemc-1.0-alpha1/src/uvmsc/macros/uvm_sequence_defines.h:80:3: note: in expansion of macro ‘UVM_DO_ON_PRI_WITH’ UVM_DO_ON_PRI_WITH(SEQ_OR_ITEM, SEQR, -1, "") ^ ahb_basic_test.h:59:11: note: in expansion of macro ‘UVM_DO_ON’ UVM_DO_ON(ahb_wr_seq,top_env->agent->ahb_sequencer_inst); Appreciate any help here. Thanks Akhila -
Hi All, I am seeing following error when I try to assign array values using inline constraints Array size's and dimensions are same and array slicing is not there ,but still I see below error. Can somebody please through light on this error? Also suggest best practices class A_seq extends uvm_sequence #(id_nfc_tx_transfer); `uvm_object_utils(A_seq) `uvm_declare_p_sequencer(tx_sequencer) bit [7:0] data_byte[`MAX_DATA:0] ; function new(string name="A_seq"); super.new(name); endfunction virtual task body(); ///Array Initialization for (int i=0;i<=5;i++) data_byte= 'hAB; `uvm_do_with(req,{ tx_bit_rate_sel ==212; tx_num_of_bytes == 5; tx_data_byte == data_byte; }) endtask endclass Error :Randomization constraint has this error, which will cause the randomize function to return 0 and no new rand values will be set: Slices of unpacked arrays are not currently supported in this context
-
I was just reading the 1800-2012 standard and I've come across something that strikes me as a possible contradiction. In section 18.8 Disabling random variables with rand_mode() it states that Later on, when an example is presented, I see the following code snipped: The comment states that packet_a.rand_mode(0) disables randomization for all variables inside the object, which is against the statement from above. Could anyone who's part of the SV committee shed some light on the issue?
