Romex Software Forum

We get feedback from support quite often, every couple of days. However, most of your questions have been commented by known users with deep usage experience with PrimoCache, so maybe that's the reason support didn't write an additional answer.

As long as you don't enable write caching, PrimoCache will be totally risk free for you to test. All you need to do is try it out and experience if it makes a change for you.

Axel Mertes wrote: most of your questions have been commented by known users with deep usage experience with PrimoCache, so maybe that's the reason support didn't write an additional answer.

Yes, I think experienced users already gave good suggestions.

support wrote: Yes, I think experienced users already gave good suggestions.

Thanks for replying. I don't mean to be a pain, but while I got some good comments from Jaga about Chrome browser cache, I did not really see a response to my two main questions about PrimoCache in this situation. Perhaps I was not clear enough:

1) Since I've used the 64 GB ExpressCard SSD already for Readyboost, and a few GB for the paging file, is it more effective to abandon that and use all of my secondary SSD for PrimoCache, or to leave some for Readyboost plus a small paging file, in addition to PrimoCache?
Note: I already did a search on this Board of all posts that mention "Readyboost", but I did not see an answer to this question. I did find comments that it's possible to use both, but not much about people's experience doing so.

2) With only 4 GB of RAM available, should I use ANY amount of this for a PrimoCache L1? Is there any benefit in taking it away from Windows? (I have x64.) My System info says that of my 4 GB -- 3.86 GB is "available". Is it even possible to use that remaining 140 MB?

Thanks,
Michael

bilateral wrote:1) Since I've used the 64 GB ExpressCard SSD already for Readyboost, and a few GB for the paging file, is it more effective to abandon that and use all of my secondary SSD for PrimoCache, or to leave some for Readyboost plus a small paging file, in addition to PrimoCache?

I think you can use a few GB for the paging file, but it is not necessary to use both PrimoCache and ReadyBoost at the same time.

bilateral wrote:2) With only 4 GB of RAM available, should I use ANY amount of this for a PrimoCache L1? Is there any benefit in taking it away from Windows? (I have x64.)

If you don't enable Defer-Write, it's ok that you don't use L1 (memory) cache since you have only 4GB ram installed.

bilateral wrote: My System info says that of my 4 GB -- 3.86 GB is "available". Is it even possible to use that remaining 140 MB?

I don't suggest to use the remaining 140MB because it is mostly used by some hardware components like video adapter.

Okay, thanks for that guidance. As I have already been very persistent, I will try to remember to get back to let everyone know how it worked out.

Michael

bilateral wrote:I have an older Toshiba laptop that is still enjoyable to use--except when it is really slow.
Limitations: Core 2 Duo; 4 Gig DDR2 Ram--maxxed out; fairly weak graphics; Hard to change HDD.

First suggestion - find out what exactly is slowing your system down (your phrasing suggests that this is an periodic, rather than constant, problem). I would suggest installing either Process Hacker or Process Explorer (or both, I'd give Process Hacker the edge since it can show combined read/write activity per process) and using these to find out what software is slowing your system down. Look in particular for sustained high CPU usage and high read/write activity as either can create a bottleneck for other processes. If you can identify a single culprit (e.g. an overly-zealous anti-virus scanner) then that might solve your problem without the need for more software or hardware.

bilateral wrote:For the ExpressCard slot I found a 64 GB SSD--not as fast as many SATA SSD's, but MUCH faster than the HDD. It is not possible to boot off this SSD--but I've found other ways to use it that are very helpful.
--I put a 32 GB ReadyBoost cache on the SSD...In addition, I put a 2nd ReadyBoost file on a 32 GB SDHC card in the SD slot. Two source files should help ReadyBoost a lot.

ReadyBoost can use multiple drives (in RAID-0 fashion) but they need to be of a similar speed for best performance. Using a much slower second disk (USB2 is constrained to around 40MB/s while a SATA-based SSD can reach 500-600MB/s) could result in slower performance.

bilateral wrote:So,...if I were to install PrimoCache:

--First of all, do you think it will help more?

PrimoCache L1 uses system memory to speed up repeated data access. This can boost performance for systems with memory to spare, but with 4GB RAM that doesn't seem to apply in your case (both Process Hacker and Process Explorer can help here since they report maximum memory utilisation in their System Information box, as "Peak Commit Charge"). Unless you have about 1GB or more to spare, it's unlikely that PrimoCache L1 will help much.

PrimoCache L2 (using an SSD to cache an HDD) is another matter. It can provide a speed boost to HDDs, but as Jaga has noted, not as much as swapping over to SSD completely. The big question in your situation is whether the HDD is the root cause of your performance issues - if so, then L2 caching should provide some benefit.

bilateral wrote:And second, would I leave the 6 GB paging file and the two 32 GB ReadyBoost files as they are now?

Leave the pagefile as it is - but ReadyBoost will duplicate the functionality of PrimoCache L2 so you would likely be better off with it disabled.

Given your emphasis on Chrome usage, another approach might be to investigate (and reduce, if possible) its memory usage (*cough* using Process Hacker/Explorer again *cough*). This might involve selectively disabling plugins, adjusting browser cache settings or looking at other browsers (I use an old pre-WebKit version of Opera which uses less than half the memory of Firefox, even with a dozen tabs open compared to a single page in Firefox - but you might want to check out Vivaldi for a more recent option).

@InquiringMind: Thanks for the helpful ideas! I will certainly look into trying out Process Hacker to take a closer look at these details you mentioned. I have used Process Explorer a bit, but can't say I've done a lot with that tool either; but I've never tried Process Hacker. I expect it will take me a while to get to the appropriate assessments you are suggesting, as I am in the middle of other projects right now. But it sounds like a great idea, so I'll try that as soon as I can.

InquiringMind wrote:...PrimoCache L2 (using an SSD to cache an HDD) is another matter. It can provide a speed boost to HDDs, but as Jaga has noted, not as much as swapping over to SSD completely. The big question in your situation is whether the HDD is the root cause of your performance issues - if so, then L2 caching should provide some benefit.

Frankly, I suspect that the video hardware, and to some extent the CPU, are the weakest links in this laptop, in the sense that by continuing to use the device, these are two hardware parts that I really can't do a thing about. I could install an SSD as the main drive, with a significant amount of work (as there is no "drive hatch" in this machine and tearing it down is particularly painful)--and since the hard drive is typically a major limiter, that might help a lot. Still, I don't see how I can ever overcome the limitations of the video and CPU hardware without junking the laptop--but then I wouldn't be asking these questions. Really, that is what led me to PrimoCache as a potentially good solution.

InquiringMind wrote:...ReadyBoost will duplicate the functionality of PrimoCache L2 so you would likely be better off with it disabled.

Okay,...and in that case, since I also have that SD card slot (which as you and I both pointed out, runs at USB2 speeds), is there any value at all in continuing to use that as a ReadyBoost device? Or even as a 2nd Primocache L2 device? Or does the same interference issue which you mentioned apply either way--that the SD card and the SSD are going to be running at very different speeds? (Or does PrimoCache take into account the differences in function between the two devices?) I am aware that Readyboost is designed to use an SD card only for caching many small files which can be moved more quickly from an SD card than consecutive small "seek-and-reads" from an HDD.

InquiringMind wrote:Given your emphasis on Chrome usage, another approach might be to investigate (and reduce, if possible) its memory usage (*cough* using Process Hacker/Explorer again *cough*). This might involve selectively disabling plugins, adjusting browser cache settings or looking at other browsers (I use an old pre-WebKit version of Opera which uses less than half the memory of Firefox, even with a dozen tabs open compared to a single page in Firefox - but you might want to check out Vivaldi for a more recent option).

I'll look into those browsers. What about the idea Jaga and I discussed of putting the Browser Cache files on the ExpressCard SSD? Do you think that would be effective, or would it just duplicate something that PrimoCache would already be doing? (It would take away some space from PrimoCache, perhaps, but would dedicate an area of the SSD directly for Chrome browser cache.)

Michael

I have not read through all of the details in this thread but thought sharing some personal experience may be helpful. I also have well enjoyed vintage system with a 4GB RAM cap running Win 10 x64. At present it has a SSD for boot drive but ran an earlier version of Primo with a 'performance' HDD for several years. Despite having access to far more powerful hardware this system is my daily driver for a variety of reasons.

Two key observations:
- Win 10 does a pretty decent job intelligently caching/compressing information in memory making the most of limited resources; it also utilizes the classic paging file far less frequently that previous Windows versions
- a 256MB L1 cache w/delayed write makes a big difference in responsiveness without taking away significant resources from the OS. Yes, there is a small risk of data loss in the event of a system crash or power loss which can be mitigated through a shorter write back delay

I offer this as a data point and have no interest in opening a debate on the merits and risks of write caching, perceived uselessness of smaller L1 caches or more convoluted solutions that involve relocating personal and/or paging files. 80/20 rule applies here...especially given the steps (eg: ReadyBoost) already taken to improve overall performance.

You have 60 days to evaluate PrimoCache. Why not give it a whirl?

bilateral wrote:...I have used Process Explorer a bit, but can't say I've done a lot with that tool either; but I've never tried Process Hacker. I expect it will take me a while to get to the appropriate assessments you are suggesting, as I am in the middle of other projects right now. But it sounds like a great idea, so I'll try that as soon as I can.

Process Hacker is very similar to Process Explorer, so you don't have to install it if you already have PE. PE should be able to give you the info you need as long as you're familiar with it.

bilateral wrote:Frankly, I suspect that the video hardware, and to some extent the CPU, are the weakest links in this laptop, in the sense that by continuing to use the device, these are two hardware parts that I really can't do a thing about.

You could consider overclocking them, but on a laptop you need to take particular care to monitor temperatures (since cooling is harder in a confined space) and be prepared to deal with shorter battery runtimes. For these reasons, it's best left until after you've tried everything else.

However I would like to ask why you consider these to be the cause? If you are getting low framerates in a game, then yes, you're likely to be correct. But no caching can help in that situation - PrimoCache can give you lower level load times but not better framerates.

Given your queries about Chrome, is it performance issues on particular websites? (video streaming, etc). In that case, the speed of your network connection (and your ISP's connectivity) is another factor to consider. In that case, filtering content (ads, Flash, Java) can greatly reduce the amount of data downloaded, as well as improving online security and privacy.

bilateral wrote:I could install an SSD as the main drive, with a significant amount of work (as there is no "drive hatch" in this machine and tearing it down is particularly painful)--and since the hard drive is typically a major limiter, that might help a lot.

Not only should it boost performance, but an SSD should usefully lengthen battery runtimes since it uses less power than an HDD. And your laptop should run cooler too.

bilateral wrote:Okay,...and in that case, since I also have that SD card slot (which as you and I both pointed out, runs at USB2 speeds), is there any value at all in continuing to use that as a ReadyBoost device?

ReadyBoost aims to tackle HDD access times by storing smaller files on flash media where faster access time can make up for lower transfer speed. If you are using PrimoCache L2, then that is far less likely to happen so ReadyBoost just adds a second (or third, once you consider Windows' own file caching) level of cache which is more likely to lessen performance.

bilateral wrote:Or even as a 2nd Primocache L2 device? Or does the same interference issue which you mentioned apply either way--that the SD card and the SSD are going to be running at very different speeds? (Or does PrimoCache take into account the differences in function between the two devices?)

For it to work, PrimoCache would need to benchmark the devices to determine transfer rates, then allocate data in reverse proportion (e.g. with 2 devices with 10 and 100MB/s transfer rates, to use the faster one for 90% of cached data). Subject to confirmation from Romex support, I don't think PrimoCache has that capability.

bilateral wrote:I'll look into those browsers. What about the idea Jaga and I discussed of putting the Browser Cache files on the ExpressCard SSD? Do you think that would be effective, or would it just duplicate something that PrimoCache would already be doing?

Try checking the size of your browser cache first. If it's only a few dozen MBs, then it's unlikely to be the choke point. Otherwise, moving the files to SSD should boost performance for page reloads, but can't do much for pages that sap CPU with heavy use of animations, Javascript, etc.