RAMdisk Test. (not a question)

So, as we discussed on Glasskuter's thread after we kinda hijacked it, I decided to test a RAMdisk from AMD.
I have one already installed on my Eurocom Scorpius from 2012 and last I checked, it was still working and
in good shape.

I downloaded the free version as per instructions and Tried to input my license but, I think either the Windows
firewall or the Hitron cable modem's firewall was not letting the connection to the Dataram servers through.
So, I disabled the Windows firewall and tried again but, to no avail. I'll check the modem tomorrow but, in the
mean time I opened a ticket to see if it could be activated remotely.

So, installed the RAMdisk and just to test it out I configured a 3Gig disk. It is working as it should. I installed a
few programs and an old game that I had on my 4TB external HDD from 2006. To my surprise the darn thing
installed and runs with no compatibility mode applied.
Also, after installing programs and the game, I restarted twice and also turned OFF the notebook to make sure
that the RAMdisk would not lose any of the programs and game installed on the RAMdisk. Restart was the same
but, the shut down was a bit slower. Didn't affect the startup at all and all of the software was still on the RAMdisk.
When clicking on the RAMdisk it started right away with no hesitation.
Now, the speed testing with Crystal Disk.
Below you will find the screens from the benchmark:



Don't know if the speed of the memory has anything to do with these numbers but, I suspect that the
faster the timing the faster the speed of the benchmark.

This is from my new 990 Pro



And these are the programs installed on the RAMdisk before and after the restart and shut down.
Nothing is lost and nothing has to be reinstalled. It just works as the one on my Scorpius (7 Ultimate)



So, there you are signed sealed and delivered.

EDIT: Should have posted these as well:

That be the case it put a hole in the thimble full of ramdisk knowledge I had. It was my understanding if power is lost, so is what's stored in ramdisk since ram is volatile. Shows what little I know.

Flashorn said:

Don't know if the speed of the memory has anything to do with these numbers but, I suspect that the
faster the timing the faster the speed of the benchmark.

Click to expand...

Actually different RAMdisk apps give different speeds. Paid ones are quite faster.
For the record, AMD Ramdisk was not updated since 2017. I use ImDisk Toolkit.

glasskuter said:

It was my understanding if power is lost, so is what's stored in ramdisk since ram is volatile.

Click to expand...

It can happen during a reset or a power loss, I consider it as an advantage though, in case of a malware infection like ransomware or SWAT raid, I can simply turn off PC and ramdisk will be cleared with no recoverable data.

P.S. There are some disadvantages, like I can not install AMD CCC, because it does not recognize Temp folder on ramdisk.

glasskuter said:

That be the case it put a hole in the thimble full of ramdisk knowledge I had. It was my understanding if power is lost, so is what's stored in ramdisk since ram is volatile. Shows what little I know.

Click to expand...

Well, my old 2012 notebook is still holding it's RAMdisk software after well, I can't remember but, the last time
I powered it up (must be sometime last year) it still had all of it's software intact. That version of the AMD RAMdisk
is about from 2014 or 2015. You also have the option of keeping the disk or removing it on shut down (power off RAMdisk).

TairikuOkami said:

Actually different RAMdisk apps give different speeds. Paid ones are quite faster.
For the record, AMD Ramdisk was not updated since 2017. I use ImDisk Toolkit.

Click to expand...

I've never seen speeds like these before. Even on my older notebook, the speeds are not
this fast. In the old notebook, the memory timings are 9 9 9 24 Corsair Vengence 2400Mhz DDR3.
This one has 4800Mhz with timing that I would have to look up. This will have an effect on the speed of the
RAMdisk right? even if it's the free version.

Update: had the license file sent to me from Dataram.
All it took was a click and it was activated.
Redid the RAMdisk with a larger 6Gig file and again had
no problem keeping the data on it after restarts and shut downs.
Re-tested with Crystal Disk and the results are as follows.

Flashorn said:

So, as we discussed on Glasskuter's thread after we kinda hijacked it...

Click to expand...

:::GRINS:::

She must be used to it by now. This thread dates back to 2023. Sorry GK. It's just that your threads are always so intriguing.

Flash, u da man. Whenever I need help with NVMe and RAMDisk I'm gonna seek you out. Seriously. You know your stuff.

Thanks for this. It gives me ideas.

Scannerman said:

This thread dates back to 2023.

Click to expand...

OH! I remember 2023 so well. (sorry, but, you left me no choice)

Whenever I need help with NVMe and RAMDisk I'm gonna seek you out

Click to expand...

I'll help in anyway I can but, you might be putting too much trust in my knowledge.

Thanks for this. It gives me ideas.

Click to expand...

Only good vibes I hope

Had to see if anyone was paying attention. Hard to believe we only have one month left to go.

Scannerman said:

Had to see if anyone was paying attention. Hard to believe we only have one month left to go.

Click to expand...

You know I always pay attention to your post.
And, to be fair, we still have 7 weeks before the end of the world.

glasskuter said:

Shows what little I know.

Click to expand...

It's not how little you know but, what you experiment with that shows you how much you now.
You are always helpful.

Flashorn said:

You know I always pay attention to your post.
And, to be fair, we still have 7 weeks before the end of the world.

Click to expand...

I had better behave myself if you're that attentive.
Well now, I wasn't aware that you complied with the conventional calendar.

You might want to try primocache. It caches at the disk block level instead of file level and you can persist it across reboots, control how much of the allocated space is for read and how much for write, and setup deferred write cache (write-back cache). I have write-back enabled, which can be a bit risky but SSD are so fast, and I have never had an issue. There are other advantages of write-back besides write performance.

This is the cache on my c-drive (WDC black nvme SN850). I also have my DDR4 RAM at 4200 MHz 18-18-24

To get the best from both worlds, I use PrimoCache in conjunction with Primo Ramdisk.

My laptop only has 16GB of G.Skill DDR4 @ 3200MHz 22-22-22-52 single rank, but Primo Ramdisk gives the option to use Dynamic Memory Management (DMM) and Compact mode so it doesn't use up more RAM space than it needs to store the actual files. Whereas the vast majority of other ramdisk softwares would cut the amount of free available RAM space in half if choosing to go for an 8GB ramdisk no matter whether the ramdisk has any files stored on it. That plus the fact that, even with DMM and Compact mode set to enabled, Primo Ramdisk still performs a whole lot faster than yours. This is on my "lowly" 599 Euros grocery store laptop that will be already 3 years old by the end of next month:

I'm currently using RAMCACHEIII (provided by ASUS) on the X570 HERO. It isn't a very extensive app but it seems to have the basics.
For now it is suffice. Not sure about later though, as I mean to switch out my NVMe at some point. So far, so good.



I do like that it allows me to go well beyond the recommended cache allocation. I ran the cache with the bar maxed out without issue but thought I would step it down a bit to arrive at "saner" looking speeds. This made little difference. I'm running the crappiest NVMe ever in this build and I'm still getting impressive speeds with this cache for obvious reasons. The speed is in the RAM.



This is not the fastest RAM in town either, but for DDR4 it does pretty good. I noticed that the DRAM information only acknowledges the base clock speed in this app. In actuality the RAM is clocked to 15-15-15 @ 3666 MHz. Latency is good and low. I'm running stock cooling and stepping the speed up higher seems to affect my CPU OC so I'm leaving it as it is. 4.2GHz is enough on the CPU for this old guy. I could probably do better with an NHD-15 or liquid cooling.