Evaluating Hardware

With modern computer hardware, the speed of the CPU (in Mhz or Ghz) and the core count alone are no longer enough to determine if a given system will perform adequately. There are many more factors to consider, such as the architecture of the CPU, its generation, the technology it is based on, its power consumption, the amount of RAM available, etc.

Benchmarks

The only way to reliably compare CPU’s is by looking at their respective benchmark results. Synthetic benchmark tests compound all CPU attributes into a single numeric score value. Quite simply put, the higher the score, the better the CPU is.

CPUBenchmark.net

PassMark Software’s CPU benchmark is a highly regarded database that contains all the information required to compare millions of different CPU models. The database can be consulted here:

Cpubenchmark.net Website

By searching this database and using the provided “Compare” tool, it is easy to determine if the hardware that you plan to buy (or already have) will meet or surpass the performance of the recommended hardware configuration.

Performance indicators

Here are a few performance indicators we can use to measure the performance:

CPU Benchmark Score

This is the score of the CPU posted on https://www.cpubenchmark.net/. Note that the posted scores are averages and may vary daily. Higher is better.

Minimum Keystroke Delay

This is the minimum time, in milliseconds, that the user needs to wait between keystrokes to avoid visible lag and missed keystrokes. Lower is better.

Methodology

This value is measured in a lab environment using a tool built into the POS software.

The input keystrokes and the POS' output are compared, and the delay is shortened until one of two things happen:

1. The POS starts "missing" keystrokes. (i.e. a key is pressed but the POS fails to register it), or;

2. There is visible lag, meaning there is a significant and clearly visible delay between the keystroke and the information appearing on the screen.

The delay between keystrokes is then increased gradually until the POS can reliably process all keystrokes without missing one and without creating significant lag, for a period of at least one full minute.

Maximum keys / second

This is the maximum theoretical number of keystrokes per second that a user can achieve without causing the POS to lag. Higher is better.

Methodology

This value is calculated from the minimum keystroke delay. For example, with a minimum keystroke delay of 100 milliseconds, we can have a theoretical maximum of 10 keystrokes per second.

Maximum keys / minute

This is the maximum theoretical number of keystrokes per minute that a user can achieve without causing the POS to lag. Higher is better.

Methodology

This figure is obtained by multiplying Max Keys / Second by 60. This figure is used to put the keystroke delay and keys per second values in perspective. Knowing that an average typist can do between 190 and 200 keystrokes per minute on a keyboard, while a "two-finger" typist does 150 to 160, this figure makes it easy to determine if the POS hardware will be able to keep up with the employees using it.

Sample hardware configurations

Multiple hardware configurations were tested in a controlled lab environment during the development of the HTML POS. These can be used as benchmarks to compare the hardware you already have or plan to buy.

System #1: Minimum

The first system tested is an All-in-One POS with the lowest hardware specification that was tested in our labs. It is also the oldest, with a manufacture date of late 2008. Still, even at this age, the hardware is slightly above the minimum requirements for the Maitre'D Legacy POS.

System #1 Conclusion

This system is not adequate for HTML POS, however it works just fine with the Maitre'D Legacy POS. If you have older hardware available which can be upgraded to Windows 10 IoT Enterprise, it can still be put to good use for a budget POS system using the Legacy POS software.

System #2: Tablet PC

This system is a Windows 10 Tablet PC, optimized for long battery life.

System #2 Conclusion

While HTML POS was usable on this system, it was a bit on the slow side. However, this is a very good wireless solution for the Legacy Maitre'D POS.

System #3: NUC form factor PC

This system is a Windows 10 NUC form factor PC with a very small footprint. The enclosure of this device measures only 5 in x 5 in, with a thickness if about 1.25 in.

System #3 Conclusion

While HTML POS was also usable on this system, the performance gain from the slightly better CPU was marginal and could not be measured. This is a very good solution for the Legacy Maitre'D POS if you need a low cost and space-efficient solution.

System #4: POS hardware

This system is an All-in-One Windows 10 POS workstation.

System #4 Conclusion

While HTML POS was also usable on this system, the performance gain from the slightly better CPU was marginal and could not be measured. This hardware is more than adequate for the Maitre'D Legacy POS, however.

System #5: High-End POS hardware

This system is a high-end, All-in-One Windows 10 POS workstation.

System #5 Conclusion

The Core i3 desktop CPU and 8Gb of RAM made all the difference with this system, doubling the HTML POS performance compared to previous systems tested. At 400 keystrokes per minute, this system now surpasses the capacity of even the most experienced typists.

System #6: Old Desktop PC

This system is a standard desktop PC manufactured in 2011, which was upgraded to Windows 10.

System #6 Conclusion

While a full-sized desktop computer does not necessarily make sense to use as a POS, this hardware was included here as a point of comparison, and to show that old hardware can still be used very efficiently. Also, similar hardware is available in Small Form Factor (SFF), which has a much smaller footprint.

System #7: High-End Desktop PC

This system is a high-end, Small Form Factor (SFF) desktop PC, running Windows 10 pro and manufactured in 2018.

System #7 Conclusion

A SFF PC such as this one has a relatively small footprint. Coupled with a touch screen, this could effectively run the entire Maitre'D Back-Office suite along with HTML POS as a local workstation without any noticeable drop in performance.