Compare your PC performance scores with Gizmo's Freeware readers


A few days ago Rob posted an article on how to get the Windows Experience Index score back. Today we are going to showcase our PC Performance Scores using another freeware utility, Novabench. It is a benchmarking utility for Windows and Mac. Novabench is a versatile and advanced aggregator than WEI that tests the performance of various components of your computer and then assigns scores to each one of them. A higher score indicates a better performance. Novabench not only rates your computer but also can be used to find the various performance bottlenecks on your system. The scores generated comes from CPU Tests that include Floating Point Operations, Integer Operations and MD5 Hashing, GPU Tests calculating 3D Frames per second, Hardware Tests measuring RAM Transfer Speeds and Drive Write Speeds.



Novabench is extremely easy to use even for inexperienced computer users. Make sure you close all the applications while running the tests to get a higher score. Click on Compare These Results Online from the results window after finishing the tests and post the link here along with your score in the comments section. Scores and Stats here shows the Highest rated PC's posted at NovaBench. I will make sure to update the Top 10 Scores of our readers in this article. 




Please rate this article: 

Your rating: None
Average: 4.2 (24 votes)


NovaBench Score

785 on a Lenovo E540 cheap laptop but NovaBench displays the incorrect GPU. It shows the integrated graphics Intel HD Graphics 4600 but the Nvidia Control Panel was used to set the default as the Nvidia GeForce GT 720M. This is a known bug that has not been resolved.

The score with the integrated graphics is, as expected, lower at 728. The reduction in 3D fps is from 403 down to 246:

NovaBench Score

score: 913
custom desktop computer, 5 years old

But look at this:

Hardware Tests (Score: 32)
- Primary Partition Capacity: 892 GB
- Drive Write Speed: 38 MB/s

Hardware Tests (Score: 21) - my computer
- Primary Partition Capacity: 165 GB
- Drive Write Speed: 91 MB/s

The benchmark is telling us that 165 GB with 91 MB/s is worse then 892 GB with 38 MB/s, but it should be the opposite.
Looks like they give too much importance to primary partition capacity; in my opinion writing speed should be more important.
If there is enough free space in the primary partition, capacity should not matter.
Also they should consider the size of the whole HD, not just the size of the first partition, since many users have a partitioned HD, and the primary partition is usually small, because it's only for the operating system.

Score 505. 4-year-old ACER Laptop.

Question: The article above says: :Novabench not only rates your computer but also can be used to find the various performance bottlenecks on your system." It's not clear to me how to do this. Other than randomly paging through the results on the NovaBench website to compare other results with mine, is there a way to see where my bottlenecks are? Thanks.

I've taken the time to explain this in some detail for the benefit of other users who have the same or similar question.

NovaBench is a component benchmark. It cannot show you actual bottlenecks because they depend on the tasks you actually use your computer for. In other words, hardware performance bottlenecks are almost always application specific. They depend upon the program you are using and what are you doing with it.

A bottleneck occurs when one or more components limits the performance of the whole system. So the demand on a component has to be sufficient to reach the threshold where throughput is throttled. For example, the performance bottleneck for processing large images in Photoshop is more likely to be RAM. A high-end game is more likely to be limited by the graphics card. A file or database server is more likely to be limited by the storage subsystem (disk and controller combination) and/or the network used.

So if you want to see where bottlenecks are occuring on your Windows system while running specific applications then you can use Windows built-in tools for finding PC performance bottlenecks. Also there are many guides like How to benchmark like a pro that will give you more information about a broad range of benchmark programs and give you assistance in using them. Such guides are primarily designed for gamers but they are also very useful for other applications.

You will probably notice that NovaBench is usually not mentioned in benchmarking guides but it has its uses.

1. Compare with similar systems

NovaBench stores the results for all sorts of systems that you can compare with your own. You already know that these User Scores exist but you didn't notice that you can search for CPU or GPUs with the same ID as yours. Go to Scores and Stats and you can search on three items:

  • "CPU/GPU Keywords"
  • "Submitted by"
  • "Score close to"

You can use NovaBench to search for a combination of your CPU or GPU and a score close to yours. That will give you an indication of how to improve your system performance by changing other components than the CPU or GPU.

While you can't use NovaBench to search for a combination of your CPU (e.g. i5-4200M) and GPU (e.g. GeForce GT 720M), you can do this search on Google, e.g. with a search term like this " i5-4200m gt 720m".

2. Compare the rank of your CPU or GPU

NovaBench summarises the results for each CPU and GPU on a ranked list of "Top CPUs" and "Top GPUs" which are other tabs of the Scores and Stats page. You can find your CPU and GPU and see what the average scores are.

You'll notice that the average CPU MHz can vary from yours simply because some CPUs are overclocked.

The GPU FPS (frames per second) statistic is more useful because it is more likely to be comparable with other benchmarks that use the same measure.

"database server is more likely to be limited by the storage subsystem"

Some database applications are memory bound, or even cpu bound. Depends on the application architecture. The one I used the most, Visual FoxPro, cached as much of the tables into RAM as they could hold, for instance, before performing operations. So just as array operations are faster than disk reads, such prefetch often means RAM limitations rather than disk ones.

Otherwise, pretty reasonable advice with "performance bottlenecks are almost always application specific". Usually it's pretty easy to consider the application types you are aiming to optimize and then buy the component power you should need to balance a system.

2040 - home built gaming rig
Intel Core i74770K, GTX770, 16GB, SSDs - I guess fair for gaming.

Comment on rating. I reviewed a lot of score of slower systems than mine and a lot rated about 3400+.. After reviewing I see novabench is really generous with scoring on RAM. I rate 2040 w/ 16GB while a slow CPU & GPU w/ 32GB RAM rates 30% to 35% higher rating. Speed for me is GPU/CPU and not really reliant on RAM when even playing intense games, I rare hit about 8 to 12GB.

Do post the link to the results so that I can update the Top 10.

My score is only 472 but it's a cheap Lenovo laptop.

1416 Score on homebuilt system 4 years ago.

Score 172:

7 year old Atom Netbook.

My desktop scored 1972, a far cry from the top (3572)

Lenovo desktop with a few mods Score 712

Mine is a self built desktop

The scores themselves are pretty irrelevant unless you are tracking your own computer's performance over a set time frame. You would hardly expect a 5 year old $500 computer to have anything close to the score of a new $3000 computer. The Novabench site lists top scores in a couple of categories and I think the top score was in the 3500 range. The one good point to their credit is that they provide a forum where the tests and results can be discussed.

The scores may have value in telling you whether to replace your 5 year old unit, or whether you can wait another year or two.

With Moore's law petering out, the later course of action is becoming the choice of many more people.

I think most users are satisfied if they get 5 good years of service from a mid range priced computer. Computers older than 5 can get a bit buggy but are still worth having if your main activities are basic internet surfing and accessing e-mail. My current laptop is now 6 so it's about time to thank it for it's good service and relegate it to a backup and testing unit.

Score: [b]1282[/b] (i3-3320/8GB/R9 270 - all at stock)

Is that a laptop or a desktop? One flaw I've found is that notebook graphics are given much lower scores compared to desktop. Mine is a R9 375M which scores a meagre 44 in graphics tests compared to yours R9 200 series which gets a 600+

Desktop, George.

Generally, I assume that any mobile CPU/GPU will have approximately half the performance of the Desktop parts that use the same/similar number+M. Intel HD graphics on notebooks are often better than their desktop equivalents though.

Desktop CPU:
Approximate Notebook equivalent (both 3rd gen i3's):

It's broadly similar to the GPU differences between desktop and M versions from nVidia and AMD/ATI.

High performance doesn't sit too well with battery life and notebooks but sometimes you can boost a mains connected notebook a little by ensuring you have the UEFI/BIOS and Windows Power settings to enable best performance, rather than power saving (which should be default when running on battery power only).

If your notebook has dual mode switchable graphics (eg. Intel HD + nVidia or AMD), it might be that only the Intel was used but the benchmark logged it as the better adapter...

Here is mine for Toshiba Portege R830. Score: 612. MC - Site Manager.

Score 858 -