A Third-Party WEI for Windows 10: Windows Experience Index

[BonzaiDuck]

This evening, after tuning up my storage system's two-tiered caching system, I started poking around Windows 10 Pro Build 1709 to look for the WEI. Apparently, there isn't any. Please tell me if I'm wrong.

To this end, I found a piece of freeware from an outfit known as "Chris-PC:"

Some freeware or shareware might make me nervous, especially if originating in a country like Romania. But the credits that include MajorGeeks seemed reassuring enough; my AV program didn't raise any alarms. I made sure to uncheck selection for installing a bundled item.

Has anyone seen this? Did I miss something in Windows 10? Or am I right that it no longer contains its own WEI component?

I suppose my system is performing pretty well. The processor is overclocked currently to 4.7 Ghz, but scores the lowest among component tests at 8.6. The graphics scores surprise me, because the stock settings for my GTX 1070 were those in effect for the test, rather than my AfterBurner OC setting. I suppose I cannot complain about the storage scores.

I invite anyone to download the software, run it and post a screenie for their system. Here is the link:

Chris-PC Win Experience Index

 

[mikeymikec]

You do realise the figures are virtually meaningless, don't you? The reason that the WEI UI was removed from later versions of Windows was very likely because there's no single definitive benchmark for CPU performance for example. Then there's the fact that the scoring was changed for each version of Windows that had the WEI UI in.

The only time I paid any attention to the figures was in the era when many graphics drivers weren't yet up to the standard required for Aero to work properly; it was a quick and easy way to know which driver to use. These days, chances are if you use the standard AHCI driver you'd still get an identical or very similar score in WEI to if you used to the most optimal driver that other benchmarks indicated a benefit with. Same goes for graphics.

If I had to guess, I'd say the WEI was created to help users get a better idea for what hardware specs were required to run Vista properly (as its requirements were way beyond what XP needed for optimal performance), but even then one would get results in the WEI that in no way reflected the difference between say running a Vista 32-bit system with 1GB RAM versus 2GB. Instead it scored how quickly the memory performed as opposed to telling the user that 1GB was nowhere near enough to run Vista optimally.

 

[JackMDS]

Maybe this can Help.

https://www.tenforums.com/tutorials/82029-get-windows-experience-index-wei-score-windows-10-a.html

 

[dave_the_nerd]

Your windows experience index is a 10. Windows 10, WEI 10, everything's 10. Relax and let Microsoft take care of you.

 

[BonzaiDuck]

Thank you all for your insights about WEI. I'm wondering if the CHRIS_PC freeware doesn't just gather the data from the same commands and options offered in the link provided by JackMDS.

Responding to mikeymikec: Yes -- that makes sense. I used to be the case that the lowest score I'd get came from the storage subsystem. My current storage scores can only be explained by my caching configuration. I guess my preoccupation with that aspect -- shown in posts over at "Storage & Memory" -- raised my curiosity enough to discover this freeware and pose my initial questions here.

And I probably need a night's sleep . . .

 

[WilliamM2]

I've used this one before:
https://winaero.com/comment.php?comment.news.220

No install at all, just unzip it. I trust that more than freeware installations.

 

[BonzaiDuck]

I've used this one before:
https://winaero.com/comment.php?comment.news.220

No install at all, just unzip it. I trust that more than freeware installations.

I'm too tired to create the screenie and post it. I optimized all of my storage, adjusted the RAM allocations for two caches. This time I got:

Processor 9.2
RAM 9.2
graphics 9.2
gaming graphics 9.9
Primary disk 9.2

I think these similar utilities utilize features already in Windows.

 

[XavierMace]

Good to know if I spend $30 on PrimoCache and waste hours of time fine tuning it, I could get an extra 0.35 on the pinnacle of useful data that is WEI.

 

[WilliamM2]

I'm too tired to create the screenie and post it. I optimized all of my storage, adjusted the RAM allocations for two caches. This time I got:

Processor 9.2
RAM 9.2
graphics 9.2
gaming graphics 9.9
Primary disk 9.2

I think these similar utilities utilize features already in Windows.

Winaero states that it just a GUI that pulls the information from Windows. As mentioned above, it is still there.

 

[BonzaiDuck]

Winaero states that it just a GUI that pulls the information from Windows. As mentioned above, it is still there.

That's what I had concluded. So these add-ins likely present comparable numbers to what we'd seen in the Win 7 WEI.

I understand that WEI doesn't give a true or exact assessment of things like CPU performance, but it might help understanding the bottlenecks in your hardware. I had always been used to seeing my overall WEI anchored to low storage scores.

Now that I've got two NVME M.2's in my system -- one for boot-system and the other for PrimoCache SSD-caching -- and I've doubled my RAM to 32 GB to allow for simultaneous RAM-caching under Primo, I wanted to see how the WEI changed. If the storage subsystem gets the same score as the overclocked Skylake processor or high-performance RAM, that's good enough for me.

 

[BonzaiDuck]

Good to know if I spend $30 on PrimoCache and waste hours of time fine tuning it, I could get an extra 0.35 on the pinnacle of useful data that is WEI.

You have to multiply the value of any investment in tweaking by four -- the number of machines I have running under Primo. I went through the 90-day trial on an old laptop that was severely bottlenecked by old SATA 2 technology. I came away spending ~ $65 +/- for the 3-PC license, which I applied immediately to the laptop.

Also, as you say so yourself, WEI may not provide a completely accurate result or anything more than a rank-ordering of features bereft of a more quantitative assessment of performance.

So as I'd said before at least a few times, the stability of the software -- its reliability -- figures highly for a speed-tuned processor, larger memory and some sort of NVME device. You wouldn't want to use software that fell down in some way per stability and reliability.

Put it another way. If one has his OS -- boot and system -- on a decent-performing NVME, you probably wouldn't waste the money on the software license. On the other hand, slower devices might make it worth investigating.

One guy used his system to do forensic analysis of disk devices from confiscated computers. He had some huge amount of RAM, and he was using a 1TB EVO NVME as a cache.

 

[XavierMace]

You have to multiply the value of any investment in tweaking by four -- the number of machines I have running under Primo.

First off, that's not how that math works unless all 4 systems have exactly the same gains (they don't) and you applied exactly the same settings. But beyond that, you haven't established the value and 0 * 4 is still zero. How much time are you saving booting/shutting down per machine? How much time are you saving on normal tasks per machine? And I mean actual numbers not the generic "it feels way faster". For example when I made the business case to an employer that switching from spindles to SSD's on new purchases only made about a 10% price difference per unit and was a minimum of a 20% efficiency increase (based purely off boot times and launch times for common apps), that made sense to them. So far all I've ever seen you establish is "it feels faster".

In addition to the software itself, you've added multiple drives and more RAM purely for the caching. Long term, you've decreased your reliability and increased your maintenance. What exactly have you actually gained besides great looking synthetic benchmarks which could also be had for free? You've basically built a computer around Primo at this point and spent god knows how many hours doing so. How long is that time investment going to need to see returns? Given your usage case, I'm going with never as I've never seen you mentioned doing anything demanding on your disk system.

You're regularly going around preaching this software to average Joe User but you've yet to show any actual proof of what Joe User is gaining in real world usage. Average Joe User is better off with a suitably sized SATA SSD. Done. End of story.

 

[VirtualLarry]

In addition to the software itself, you've added multiple drives and more RAM purely for the caching. Long term, you've decreased your reliability and increased your maintenance. What exactly have you actually gained besides great looking synthetic benchmarks which could also be had for free? You've basically built a computer around Primo at this point and spent god knows how many hours doing so. How long is that time investment going to need to see returns? Given your usage case, I'm going with never as I've never seen you mentioned doing anything demanding on your disk system.

You're regularly going around preaching this software to average Joe User but you've yet to show any actual proof of what Joe User is gaining in real world usage. Average Joe User is better off with a suitably sized SATA SSD. Done. End of story.

Oh, be nice to @BonzaiDuck . He's just a "tinkerer" like most of us. I don't think that he really needs an ROI case to make for his PrimoCache. For most users? Probably overkill, or at least, questionable advantages, but hey, that hasn't stopped millions of users from enabled the RAM cache in Samsung's Magician software either, even if it also has questionable tangible benefit to most people's workflows.

Edit: I think that we all benefit from BonzaiDuck's experience with the software. Let users make their own decisions about it, I don't think BonzaiDuck is trying to mis-lead anyone, at least not intentionally. Though, you are right to be objective and skeptical about performance claims, just like people occasionally posting benchmarks in here of their Samsung SSD, and of course, they installed Magician with all the options (RAM caching enabled).

Edit: What about AMD's advertising ("pushing") their AMD-branded "FuzeDrive" tiered/RAM/SSD caching program (not free)? Some enthusiasts will spend extra, just to gain a few minute FPS or a few seconds less delay doing some task on their PC.

Why else would be buy faster grades of RAM? The ROI on that is a pretty-steep decline, but enthusiasts do it anyways, just for bragging rights.

 

[BonzaiDuck]

Well, thanks, Larry. But he notes something I'd always tried to deny as an intention: I'm "promoting" the software with my unwarranted enthusiasm.

If we were running an IT department for a small company, we wouldn't be building systems like this one. Usually I take several months to make decisions about basic parts, cooling strategy and so on. I dumped the idea of water-cooling (at more expense) -- but paid a little extra for a delidded and binned CPU. I wasn't planning on installing an NVME until the last minute of choosing and ordering parts, which coincided with a fall-season hype for the 960 Pro and EVO in 2016.

But the Primo-Cache was already a useful boost to my old C2D SATA2 laptop. So it became an "experiment" for my latest systems. At first, a $130 experiment with a 250GB EVO NVME.

It just "is what it is." It gives me "good benchmarks;" it seems to be extremely stable and reliable software. The only delays in loading software or data seem to be built-in to the software itself, as with splash screens of programming credits that fade in and fade out when loading games or ending them.

Also, I freely acknowledge that "PC boot-up time" is actually increased when the RAM-cache feature that reloads the previous-session's cache into RAM is enabled. But boot time isn't a priority with me.

There is also another lesser advantage and you can quantitatively assess it as you may -- to translate into "dollars." Cached electromechanical drives suffer less wear and tear. How much? Probably depends on the usage profile.

Also -- Larry mentioned Samsung RAPID. Again, that belongs in a category with Intel ISRT, since they are both faux-proprietary caching configurations. In fact, Marvell's Hyper-Duo is proprietary to the Marvell controllers. With the former two, you would need to use two different controllers to make those features work simultaneously, and then there might be issues about which cached drive you would boot from. The only reason for the different controllers: you need AHCI mode for RAPID, and you need RAID mode for ISRT.

The $30 software has no such restrictions. I've even run RAPID alongside Primo without conflicts. But why do that?

If I made hardware choices specifically for the Primo software imperative, it came after I added the 960 EVO back to the system after replacing it with a PRO. And that effectively limits my GTX1070 to PCIE x8 -- but there's no performance loss there, either.

If one wanted to be a hardware purist and double the NVME speed, you could get a certain type of PCIE card for four NVME M.2 devices to be used in an x8 slot. I'm thinking you could create a RAID0 configuration, and the motherboard would need to implement PCIE bifurcation. There ARE Z170 boards which do that, but ASUS doesn't have any plans to include the feature in a BIOS upgrade.for my board.

I'm sure people could be perfectly happy with an SATA SSD. When I bought my first, it was a big deal. But I was already using a small one to SRT cache a VelociRaptor.

Knowing then what I know now, I would've done better to just buy a WD Blue drive to use in the caching.