The FAQ section contains the frequently asked questions and answers that are applicable for all student team. Please watch this page for updates. To submit questions about the Student Cluster Competition, please fill out the web form below.
Q1. Where can I stay in Hamburg? Where can I find a list for recommended hotel?
Please refer to the ISC'12 web for complete information. The direct link is: http://www.isc-events.com/isc12/hotel.html.
Q2. Will we be allowed to use hardware or software that has not yet been publicly or commercially released?
You are free to use any kind of hardware or software, as long as the power for the hardware is from the given budget – you are not allowed to generate power in any way or form on top of the given power budget.
Q3. Is modification of the source code , operating system or hardware allowed?
You are allow to do any modifications, again, as long as you are not using more than the given power budget.
Q4. What is all included in the 3000 Watt power limit? Would a desktop pc controlling the cluster or a nagios monitoring box be included?
A laptop/desktop machine that does system monitoring is ok not to include in the power budget. The power budget covers all of the equipment that is actively used for the applications and benchmarking (for example servers. Monitors for example is not part of the budget.
However, if you plan on using a system provisioning software with management, that is responsible for distributing the software images/packages to the compute nodes. That will count as a part of the cluster and should count as a part of the power budget.
The rule is, if it has any direct impact on the application performance, and we need to account that for the power budget.
One exception: Using external/stored power source is not allowed. Such as using battery backup (for instance, Uninterruptible Power Supply or UPS) is not allowed.
Examples of power usage that counts toward power budget:
- network switches
- system provision nodes
- laptop/desktop running system monitoring software for provisioning system images.
- File system server (such as NFS server)
- Fans, and other method of cooling (such as liquid cooling)
- laptop/desktop running system monitoring software for the power usage.
Examples of power usage that does not count towards power budget:
- laptop/desktop running system monitoring software for the power usage.
Q5. How is the power provided? One socket with 3000 Watt or two sockets with 1500 Watt each?
Right now the plan is to use 1 power distribution unit (PDU) that can delivers 3000W. Tentatively we plan on using the APC AP7821 PDU for each team to monitor the power usage through its Ethernet interface with SNMP. The TypeF wall plug from the conference floor can supports up to 16A and 3000W power. The team will use the power cap at ~13A. At the competition booth, there will be an adapter from the typeF wall power to the C20 socket for the input of the PDU.
The AP7821 PDU will provide output for up to 8 devices. If more devices are needed to be plugged into the PDU to account for the power budget, the team will need to prepare and bring enough splitter cables (IEC C14 to C13 Splitter power cables) to power the additional devices. If the type of the plugs used for the machines are not taking C13, the teams will be responsible for bringing suitable power cables.
Q6. The two secret benchmarks are presented on the first and second day, when do we have to hand in the results for these benchmarks?
There will be one secret benchmark on the first day, and another given on the second day. You will be provided with the needed information per application at the morning of each of the days.
Q7. Do we need any special version of any program to run the benchmarks. Can we use the DEISA benchmark suite?
You may try to run and discover any type of benchmark dataset and suites that are available for your practice purposes. As for preparation, We expect teams to build the applications and run a few benchmarks to get familiar with the application and dataset to discover the characteristics of an application and datasets to their own advantages.
Q8. What is the benchmark metric? The total runtime of a program or only the runtime of a part of a benchmark, e.g. the solver (see also 4th question regarding OpenFOAM)?
The (HPL/HPCC) benchmarks will be graded according to their performance.
The metric for grading for applications will be based on their runtime. In other words, all of the application runs will also be graded for their runtime performance within the power budget.
For OpenFOAM, the metric will be the runtime for the solver for each of the dataset given.
There are many areas (system hardware, software compilation tuning, process placement, MPI tuning, plus more) that can affect the application runtimes. Deciding on the right tuning factors on those areas could improve the runtime performance.
Q9. What is the order of magnitude of the benchmarks' runtime? Minutes or hours?
It will be a combination
Q10. In which form are the benchmark inputs provided?
Tentatively speaking, the input files will be given in a USB drive to each team at the beginning of each of the days, along with the mystery application of the day and any special instructions on running the applications dataset.
Q11. How are the results compared? What kind of tolerance is there. When is a result considered unacceptable or correct? Is it for instance allowed to calculate some parts in single precision?
Specific instructions on what floating point precision the application needs to run will be given along at the time when applications are handed out at the beginning of the day.
Q12. Are the cases to be run with OpenFoam during the competition known? If yes, what are they?
No, the benchmark cases will not be made public until the competition begins. Teams are welcome to familiar themselves with any public dataset available.
Q13. Is it possible to modify the System/fvSolution file to use different solvers? (e.g. if a third party library uses a different name for the CG solver ) . Is it possible to modify other configuration files like System/controlDict System/decomposeParDict?
The modification for decomposeParDict is allowed in order to change the number of processes to run. For controlDict since it affects the simulation start/end times and the deltas, it basically controls the output of the simulation, the modification for that is forbidden.
For changing the System/fwSolution for using different solvers, we have not yet decided on it yet. I tend to think that since the dataset defines which solver to use, the results from a different solver might be very different from the acceptable result using the original solver defined by the dataset.
Q14. Is the preparation of the case also counted in the runtime ? i.e. runtime of blockMesh, decomposePar and so on.
No, only the runtime (for solver) will be graded as runtime for the final results. Preparation time will not be counted.
Q15. What is measured to assess the performance of OpenFOAM?
The metric for assess the performance of OpenFOAM will be time it takes to complete running each dataset given. Only the time used in solver will be graded, the preparation time (such as time used in creating the mesh, domain decomposition) will not be counted or graded towards the final runtime for submission.
Q16. Is it correct to suppose that the students do not have to pay the registration for the ISC? What about the supervisor?
Students who participate at the Student Cluster Challenge will not need to pay for the registration for ISC. Supervisor will not have to pay.
Q17. Is it allowed to change the BIOS settings (either change in the BIOS menu, in EFI shell or in OS-level tools) and reboot in the systems during the competition?
No changes in the BIOS are allowed after the competition has started. The systems are to be powered on at all times and no rebooting. System reboots can only reboot when the system is hung or it has encountered major failures.
Q18. I was on the site not specified the size of the input files are for the applications. Would be good to know, however, so you can choose the size of the disks accordingly? Can you make as an estimate?
The input files will be handed to each team using a storage media, for example USB(s) or DVD ROM(s), before the competition starts on each day. For some applications, even the input files may be small, there are steps involved, such as domain decomposition, which will create additional data files for each CPU to process. Thus generate much more space than just the input files that are given. Teams will need to be familiar with the benchmark applications to understand the typical data set and the runtime IO requirements in order to come up with a good enough disk storage solution.
Q19. What exactly happens if you take exceeds the 3000W-line? If we disqualified now? Or do you get negative points? Or you can simply try again?
We will collect power consumption from the PDU to monitor the power usage by the teams. As soon as the power consumption is over 13A, the overload warning light on the PDU will be illuminated to notify the team has exceeded the power budget. The SCC supervisors will be notified too that the power exceeded the limit. Points will be deducted each time the power budget is exceeded.
If the power consumption get well beyond the 13A limit, Bad Things™ will happen and the circuit breaker will trip and you will lose power to the systems. The team will also lose valuable system time to run the dataset while the circuit breaker is being reset, so the team will need to control the power draw throughout the competition.
Q20. What is provided in the competition place? Is there a LED monitor or we shall take our own?The AP7821 PDU will provide output for up to 8 devices. If more devices are needed to be plugged into the PDU to account for the power budget, can we prepare a power strip which will be plugged into the PDU?
There will be 2 Type F power outlets (rated at 16A each) will be provided for each team. Only one PDU will be given to each team which will be plug into 1 of the power outlets. The other Type F power outlet will have a power strip. Each team will be responsible for bringing in accessories, such as power strip (that takes the Type F plug), power cables for electric equipments, computer monitors, splitter cables (with proper plugs), etc. If you need to connect more devices, the team will need to bring in additional power strip. Please note that computer monitor can be excluded from the 3000W power budget.
Q21. About the 3000W power limit, we want to know the detail rules, i.e. when the cluster's power is over 3000W, would the power be turn down immediately or there is a little time before the power turned down? Points will be deducted each time the power budget is exceeded how many each time ?
The AP7821 PDU will be set to trigger a warning light on the PDU (at 13A). As soon as the cluster power exceeds 3000W (or 13A) during the actual competition, the team will see the overload warning light on the PDU will be illuminated to notify the team has exceeded the power budget. The SCC supervisors will be notified too that the power exceeded the limit. Points will be deducted each time the power budget is exceeded.
SCC supervisors may disqualify one or more run submissions if they observe that the power consumption is beyond the limit for an extended period of time. The idea is for teams NOT to exceed the 3000W (or 13A) limit. The team needs to control the power draw to 13A or less.
If the power consumption reaches around 16A, Bad Things™ will happen and the circuit breaker will trip and you will lose power to the systems. The team will also lose valuable system time to run the dataset while the circuit breaker is being reset, so the team will need to control the power draw throughout the competition.