Search the Community

Showing results for tags 'UVM'.

More search options

  • Search By Tags

    Type tags separated by commas.
  • Search By Author

Content Type


  • Accellera Systems Initiative
    • Information
    • Announcements
    • In the News
  • SystemC
    • SystemC Language
    • SystemC AMS (Analog/Mixed-Signal)
    • SystemC TLM (Transaction-level Modeling)
    • SystemC Synthesizable Subset v1.4 Public Review
    • SystemC Verification (UVM-SystemC, SCV)
  • UVM (Universal Verification Methodology)
    • UVM 1.2 Public Review
    • Methodology and BCL Forum
    • UVM SystemVerilog Discussions
    • Simulator Specific Issues
    • UVM Commercial Announcements
  • Portable Stimulus
    • Portable Stimulus Discussion
    • IP-XACT Discussion
  • IEEE 1735/IP Encryption
    • IEEE 1735/IP Encryption Discussion
  • OCP (Open Core Protocol)
  • UCIS (Unified Coverage Interoperability Standard)
  • Commercial Announcements
    • Announcements


  • SystemC
  • UVM
  • UCIS
  • IEEE 1735/IP Encryption


  • Community Calendar

Found 68 results

  1. Hello All, I am working on SystemC-UVM based testbench. I have created UVM based testbench using UVM code generator. Here I am using three different agents in UVM generator which I have defined as a part of the configuration file which was further being provided to UVM code generator to generate test bench skeleton. And here, in of the agent driver, the code snippet is as: // Drive the inputs of the DUT UVM_INFO(this->name(),"Driving transaction:",0); req.print(); // TODO put your code here But when I try to read the value in my test-case, then I see an error as: And this error is observed inside print() function, once I comment the line "req.print()" from the above snippet the error is no more observed and my test-case runs fine.
  2. Hi , To whom may correspond I think there is some kind of error in the UVM 1.1d register model. I have been experimenting with the UVM register model and i have seen the following code in uvm_reg_map.svh task uvm_reg_map::do_bus_write (uvm_reg_item rw, uvm_sequencer_base sequencer, uvm_reg_adapter adapter); uvm_reg_addr_t addrs[$]; uvm_reg_map system_map = get_root_map(); int unsigned bus_width = get_n_bytes(); uvm_reg_byte_en_t byte_en = -1; uvm_reg_map_info map_info; int n_bits; int lsb; int skip; int unsigned curr_byte; int n_access_extra, n_access; int n_bits_init; Xget_bus_infoX(rw, map_info, n_bits_init, lsb, skip); addrs=map_info.addr; // if a memory, adjust addresses based on offset if (rw.element_kind == UVM_MEM) foreach (addrs[i]) addrs[i] = addrs[i] + map_info.mem_range.stride * rw.offset; foreach (rw.value[val_idx]) begin: foreach_value uvm_reg_data_t value = rw.value[val_idx]; /* calculate byte_enables */ if (rw.element_kind == UVM_FIELD) begin int temp_be; int idx; n_access_extra = lsb%(bus_width*8); n_access = n_access_extra + n_bits_init; temp_be = n_access_extra; value = value << n_access_extra; while(temp_be >= 8) begin byte_en[idx++] = 0; temp_be -= 8; end temp_be += n_bits_init; while(temp_be > 0) begin byte_en[idx++] = 1; temp_be -= 8; end byte_en &= (1<<idx)-1; for (int i=0; i<skip; i++) void'(addrs.pop_front()); while (addrs.size() > (n_bits_init/(bus_width*8) + 1)) void'(addrs.pop_back()); end curr_byte=0; n_bits= n_bits_init; The code continues but the interesting part is already there. Lets assume we have a register with 4 bytes and 1byte per address granularity (byte_addressing). Now, we do a FIELD access of 8bits length (the first byte of a register). The field is configured "individual_accessible, so UVM should only access that FIELD. The reg2bus should generate that byte request to be written. In other words, the vector "addrs" should have only one byte address. Going to the code, i see that initially the addrs has the 4 address ( the whole register) and when it comes to the "if (rw.element_kind == UVM_FIELD) begin" and it will pop_back()/remove all the exceeding bytes that doesn't need to complete "n_bits_init" of the field access. The problem is here: UVM has //while (addrs.size() > (n_bits_init/(bus_width*8) + 1)) and i think it should be while (addrs.size() > ((n_bits_init-1)/(bus_width*8) + 1)) //ejonalv possible error in UVM? check That is because in case we want to write 8 bits, it will calculate 8/8+1=2 address in the UVM version, but in fact it should require only 1 address. This is of course applicable for the READ variation. Did i misunderstand something? It is very hard to go through the register model without proper documentation in the code. I am looking forward your answer Best Regards Jonathan
  3. Hi, The purpose of this discussion is to understand different possibilities by which the simulation performance, memory usage can be increased. Scoreboard as we understand needs the data/packets/frames etc to be stored/buffered to do a comparison with the actual data out. This works out fine when we have small sized array ranging from few bytes to few thousands of bytes.. However imagine, if we have 100 thousands of bytes and above and there is a need to store them and lets say multiple of such lanes/flows , then this would take a hit on the simulation performance. And this gets worse if we have to reuse and port it to a subsytem/chip level simulations.. With this as the background, i am looking at alternate approach for scoreboards.. In a way is possible avoid scoreboard and check the data as it comes without the need to store the expected/input data. Approach that can be thought of : 1. Generating incremental data pattern and check at the output .. (Comes again with issues like aliasing etc.. and hence need to have longer patterns.) 2. Generation of PRBS stream as the payload of the frame.. and then have a PRBS checker at the output.. and see it remains locked.. (Debug would get worse with this.. if there is a mismatch) 3. Having a byte scoreboard.. something like that.. (Not sure how feasible this is). There are many challenges even with these approach.. I am just wondering if someone has already experienced such cases and what kind of challenges are thrown out.. Appreciate any input on this. Regards, svuvmuser !
  4. This is a online training batch taught by experts working in Semiconductor Industry. Course starts with System Verilog and covers advanced level UVM. Course schedule is for 2.5 hours on Sat and Sun for 7 weeks. Graduates and working professionals can benefit from this training. Next batch starting on July 11th, 2015! For more details please email Course schedule is for 2.5 hours on Sat and Sun for 7 weeks. Graduates and working professionals can benefit from this training. For more details please email
  5. Hi why does phase.find_by_name(.name(uvm_main_phase::get().get_name()), .stay_in_scope(0)); from run_phase and uvm_pkg::uvm_phase run_phase = phase.find_by_name(.name(uvm_run_phase::get().get_name()), .stay_in_scope(0)); from main_phase return null. Is this intentional? I have created a sample example to check, or you can use the attached file as well. test.svh
  6. Hi, I have a doubt regarding an interface "X" being defined as : Driver code of the VIP: *********************************************** protected virtual X ChipIf; function void connect phase(uvm_phase phase); Y vif_container; super.connect_phase(phase); // Assigning virtual interface assert(uvm_config_object::get(this, "","V_IF_CHIP", temp)); $cast(vif_container, temp); ChipIf = vif_container.get_v_if(); endfunction : connect_phase ********************************************************** From the tb_top of the SoC , where the above VIP pkg is imported along with uvm pkg:*,please let me know how can I "set" the virtual ChipIf?
  7. The VCS implementation of uvm_reg_bit_bash_seq UVM register bit bash sequence performs a model.reset() in the sequence body, before starting the core do_block() task. Due to this reset, any configurations made to the DUT before starting the bit bash sequence is lost in the mirror model, while the DUT still has the configuration intact. This is causing failures during the bit-bash process, resulting in a test fail. There is no knob to override the reset functionality, nor can I extend the sequence and bypass the reset. Any thoughts on this? Any work around for this? ~Chethan
  8. I'm trying out the example for UVM-Connect 2.3 and I can't get a successful compile. The error message is about the "undefined reference to `m__uvm_report_dpi'. I'm using: GCC 4.5.2 on CentOS 5.11 VCS 2015.09-SP2-3 SystemC 2.3.1 SCV 2.0.0 UVM 1.2 UVMC 2.3.0 Appreciate all the help!
  9. The cook book from Mentor tells following and in another thread, the moderator also suggested against using the sub phases of run. However in one of my projects, I do find the need for using them (and infact we had an internal implemention of something similar in our previous OVM version). Are there any thing happening on this front? Is there a risk in using the sub phases if some of that changes in a future version? "The Accellera UVM committee is still developing the use models and APIs for new UVM phasing as it relates to sequences and transactors. In the mean time, our recommendation is to wait until that work is done and the API is stable. There are a number of future articles in this section which will be available here at that time, and which will describe our recommendations for using this technology. These include: How to make your testbench phase aware [Not yet available] How to manage sequences in the context of phasing [Not yet available] How to design reusable transactors that work in the context of phasing [Not yet available] "
  10. Let's say I have the following DUT. The UVM environment contains a chain of models/predictors. Input data flows down this chain and generates the expected CHIP output, which is compared to actual. Pros: verifies top-level functionality. Cons: Does not verify block level functionality. A good start, but I'd like to also verify the blocks in a system setting. So, I create block-level environments, then reuse them at the top level. Awesome, but wait a minute. I still need the top-level verification (Input-to-Output) like in the first example. However, all 3 of my block predictors are being used in their corresponding environments' scoreboards, hooked up to the RTL, via agents. How does one do both? Surely I'm not supposed to instantiate duplicate copies of my block level predictors to create the end-to-end model chain...
  11. Hi, We are using snps ralgen to generate the regmodel. It appears that the ralgen creates only the default map. We would like to have 2 maps for 2 separated if masters. Is there an online example for such case? A post gen script can do one of the following 2 options: 1. add another instance of uvm_reg_map and copy/clone the ready map after finished build 2. add another instance of uvm_reg_map, and duplicate any map1.add_reg and map1.add_submap to map2 Which is preferable? Thanks! Elihai
  12. I see uvm_sequencer_base::wait_for_grant (UVM 1.1d) is a virtual task but accesses a local int g_request_id - is this not a bad coding style? If I were to override this virtual method for debug with much of the code intact tool throws an error for his local bar in a derived SQR class. I extended a SQR class and copied all the code for wait_for_grant and started tweaking - couldn't proceed with that debug due to this member being local. Should it be protected instead of local? Thanks Srini
  13. Is it possible to do a uvm_config_db::set() for an object of derived class type using the base class handle and later do a uvm_config_db::get() of same object using the derived class handle Since a base class handle can be used to point to a derived class object and later typecase, I thought this would work, but doesn't seem to be so. Any help will be appreciated? Here is an example class BaseA ; endclass class DerivedA extends BaseA endclass 1) Set config_db DerivedA a1; a1 = DerivedA::create() uvm_config_db#(DerivedA)::set(this, "" , "myobj", a1); 2) Get config_db BaseA a1; DerivedA a2; uvm_config_db#(BaseA)::get(this, "" , "myobj", a1); $cast(a2, a1);
  14. I need to have two uvm_tlm_target_socket in a class and I need to do different set of things with the data received via two sockets. I was thinking if it is possible to have two implementation of b_transport task. I am aware of how we can have multiple analysis port imp and use uvm_analysis_imp_decl(_something) and we can have a "write_something" implementation. Is something similar available for uvm_tlm_target_socket? Regards, Gautam
  15. Hi, I notice something interesting in the build_phase order of uvm_component. The uvm_component at the same level are build in the alphabetatical order of the instance name. I expect the build order would be the same as the order I call the factory create function. I am wondering is the alphabetatical build order an intended feature of UVM or just some artifact of the implementation? I couldn't find any reference to this behavior in the UVM user guide or the class reference library. Thanks. Horace
  16. I'm trying to run uvm-systemc on macosx. Link to download: In the install flow, ../configure works fine, but on make i get this error: Making all in macros CCLD libmacros.laar: no archive members specifiedusage: ar -d [-TLsv] archive file ... ar -m [-TLsv] archive file ... ar -m [-abiTLsv] position archive file ... ar -p [-TLsv] archive [file ...] ar -q [-cTLsv] archive file ... ar -r [-cuTLsv] archive file ... ar -r [-abciuTLsv] position archive file ... ar -t [-TLsv] archive [file ...] ar -x [-ouTLsv] archive [file ...]make[4]: *** [] Error 1make[3]: *** [all-recursive] Error 1make[2]: *** [all-recursive] Error 1make[1]: *** [all] Error 2make: *** [all-recursive] Error 1 I've looked online, and it seems that it is a makefile problem. However the uvm-systemc makefile is way to complex for me to comprehend where the issue could reside. Any ideas on why is this happening? Thanks
  17. In case of UVM ,config_db can be accessed in any component or object which helps mainly the dynamic creation of component/models.Whether any option is present in SYSTEM C ?
  18. Hi, I have a doubt about requirement of raise/drop_objection. Why does a compiler need objections in run_phase? Why can it not just wait for time given like 100ns as following example? Ex. task run_phase(uvm_phase phase); //phase.raise_objection(this); #100ns; //phase.drop_objection(this); endtask
  19. Hi all, I'm new with UVM and I came across a problem. I am working on an AXI RD VIP using UVM and I have the following issue. In the data_phase (the data channel driving) from the MASTER driver, I need to drive the RREADY signal. There are 3 handshake types described in the protocol specifications: valid before ready, ready before valid and ready and valid at the same time. In case of valid before ready, I want to wait a certain number of clock cycles ( delay ) from the time that RVALID asserted and then to assert the RREADY signal. From my understanding, all the delay information should come from the transaction(sequence_item). The problem is that I generate a new transaction (sequence_item) with get_next_item only when I drive the address channel (in address_phase); The data_phase works in parallel with the addr_phase and its independent of the address phase. Also the data_phase needs multiple delay values (one for each data received from DUT, ex. arlen+1) while I only generate one transaction that contains the address channel informations. Code example: task run_phase(uvm_phase phase) fork forever begin ... seq_item_port.get_next_item(req); address_phase(req); seq_item_port.item_done(); end forever begin data_phase(); end endtask : run_phase task address_phase(axi_item item); // Drive the address channel ... endtask : address_phase task data_phase(); // Wait for the rvalid signal while(!vif.rvalid) @(posedge vif.clk); // Insert delay between rvalid assertion and rready assertion repeat(<problem!!!!>) @(posedge vif.clk); rready = 1; ... endtask : data_phase I don't know what variable (sequence_item variable) to set in the <problem!!!> Can anyone give me an advice regarding this problem. I repeat, I want all timing related data to be set from the sequence_item Regards, Adrian
  20. Hello there, As per Mentor's UVM guidelines 5.2 [1], reset_phase() will be obsolete in future releases. During DVCon 2014, Cadence recommends to use run_phases() on slide 5 of [2]. With the release of UVM-1.2, I believed that the sub-phases of run_phase are now stable and clean. Now with the recommendations above, it seems that the it's better to stick with run_phase() itself. As UVM delelopers, what are your views about it ? What do you recommend ? My intention is not to start a flame war of any kind. But to understand which route to opt in order that most of my UVM code would be compatible with UVM 1.3. All in all, it appears that there is a miscommunication on the web. [1]: [2]: regards, Chitlesh
  21. Hi All, I have a sequence sending a created and randomized item using `uvm_send. The driver receives an item using try_next_item. Upon receiving, it drives the item and calls item_done. Using debug message after item_done, I can clearly see that item_done is called and returned but `uvm_send in sequence is still blocked and not doing forward to send next item (it implements a loop). Can anyone help me with possible reasons why `uvm_send would not return even when driver has called item_done and come out of item_done. Thanks in advance! Ninad
  22. Hi all, I'm doing verification for an PHY between SPI master and a memory chip. I make two agents one for master to transfer the request, one mimics the memory slave to reply. PHY will be hooked up to two interfaces that of SPI Master and Memory. During sending request and reply data, Chip Select Pin (in SPI interface) must go low to enable the transaction. But I don't know how to control this pin when It sends the reply from memory. Because this pin is not an interface of memory slave agent. Could anyone give me some advice? Could I use phases to control the env that has different agents? Thank you, Nhat
  23. Hello, I'm trying to implement an AXI Slave VIP and have few questions regarding the implementation. In this case, the DUT is the master. The AXI Slave checks the interface for valid read /write signals and performs a read/write operation from a memory model. It returns back the write response/read data back to the DUT. 1. Since this is a slave VIP , do I need a slave sequence which runs forever sending transactions to the driver ? This is similar to the UVM example where the monitor and sequencer are connected by an analysis port and the sequence calls the peek function to check if a valid transaction is available from the monitor. (OR) 2. Can I skip the sequence/sequencer part and just connect my monitor and driver using an analysis port and pass on the observed transaction from the monitor to the driver for further action ? (OR) 3. Im thinking of a 3rd alternative of just using the monitor to the observe the interface and drive back the write response/ read data back using the monitor itself and leave the driver empty. Please let me know your valuable thoughts and suggestions. Thanks, Madhu
  24. Hi, As per my understanding, connect_phase does not start until all build_phase do not complete. How is this mechanism controlled? I did not find anything in uvm reference manual about this. Please let me know if there is any. Thanks.
  25. The Easier UVM Coding Guidelines

    Version 2016-06-24


    This document is a printable version of the Easier UVM Coding Guidelines from Doulos. You are free to use these guidelines directly, to merge them into your own company-specific UVM coding guidelines, or merely to borrow some of the ideas. These coding guidelines are offered by Doulos for the benefit of the UVM community. They are not officially endorsed by Accellera.