• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Should i buy primocache?

Status
Not open for further replies.

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Update with the corrected settings - I made sure to preload/reuse the cache this time, not make a new one each boot on the read-only (which accidentally had write buffering on too) and set it to preload. I mostly play the same games anyway, so why not.

There's a slider to set the amount used for read/write vs auto, i highly recommend 100% read for a games drive, it's not like write cache is going to help a steam game updating, but it will help with temporary files from an OS/browser and downloads (especially large fragmented ones)

C:
50% reduction today
1679199107817.png

1679199234154.png



D:
80% hit rate on the games
1679199135688.png


1679199274430.png
 
Joined
Dec 12, 2020
Messages
1,755 (1.37/day)
Wow a 32 GiB read RAM cache.

I've been using a Sk-hynix P31 gold 1 TiB NVME M.2 SSD to cache my Samsung 870 QVO 4 TiB SSD. The performance of the P31 Gold is much better across the board than the 870 QVO and the one time the 602 GiB primocache L2 R/W cache got dumped the stuttering in Metro:Exodus was instantly noticeable.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Wow a 32 GiB read RAM cache.
Windows decided to add it's own 20GB of caching, so i've got 52GB of read cache on a 64GB system since the games and OS use less. It's kinda silly, but *also the reason i wanted the RAM in the first place
(he windows caching doesnt remember from a previous boot, and yeets itself if something needs it - while primocache hoards it like a dragon with RAM sticks for the shiny gold contacts

(*That, and the sale when i bought it. $400 Au for 2x32GB right before covid janked prices up)

Interesting SSD combo, i'm surprised that the QVO wasn't enough for the game - it's not like their read speeds are terrible.

PC's only been on 30 minutes and edge/TPU exclusively, but look at this nuts write reduction
If i had a QVO as an OS drive, i'd be running a long-ass delayed write on it.
1679377683136.png
 
Joined
Feb 1, 2019
Messages
2,765 (1.41/day)
Location
UK, Leicester
System Name Main PC
Processor 13700k
Motherboard Asrock Z690 Steel Legend D4 - Bios 13.02
Cooling Noctua NH-D15S
Memory 32 Gig 3200CL14
Video Card(s) 3080 RTX FE 10G
Storage 1TB 980 PRO (OS, games), 2TB SN850X (games), 2TB DC P4600 (work), 2x 3TB WD Red, 2x 4TB WD Red
Display(s) LG 27GL850
Case Fractal Define R4
Audio Device(s) Asus Xonar D2X
Power Supply Antec HCG 750 Gold
Software Windows 10 21H2 LTSC
Quite possible - i've not kept up on the full details of all the file systems other OS's are using.

Longer term testing results:

I managed to get a 95% hitrate on the games drive cache playing the same games repeatedly, but i'd left it able to do write caching by accident - so a steam update ate the cache. sad.
Adjusted settings, but that reset the stats.
Load times were already fast, but now it's things like intro videos and loading animations holding me up, not the hardware itself.


My C: drive was updated to a 60 second cache and forced write-only, and it's info's still good to use.
View attachment 288290
How to interpret this:
Cache is write only, 60 second delay. Despite this it's got a 37% *read* rate where it wanted to read and re-use that data before it was even written the first time (temporary files in browser, is my guess)

There are some deferred blocks (~5%) right now as i'm torrenting a linux ISO (Latest Mint), but the key takeaway is that the total writes are at 65% (70%, once that download finishes)
30% extra lifespan on my 970PRO means it will outlive me.

Oh and the windows xbox overlay confirms it doesnt cause issues with directstorage - this should cache for that too in the future, resulting in some stupidly fast load times
(3x NVME drives, two have partitions)
View attachment 288292


edit: it seems torrents *really* benefit. I assume because smaller pieces get stitched into bigger ones, the longer the delay.
View attachment 288309
(and then moving the completed file to the other drive while still in the cache, boosted the read-hit rate. Interesting for use with a mech drive that moves a completed file elsewhere.)
Yeah torrents write small bits of data at a time, so what you experienced is what I would expect.

I can see myself replacing my static winramtech ramdisk with this, the problem now with the static ram disk is that browsers have moved on a lot in how they write data (no longer that effective to just dump temporary inherent files on ramdisk) and this is clearly much more intelligent able to help with OS logs and so on.
 
Joined
Dec 12, 2020
Messages
1,755 (1.37/day)
Windows decided to add it's own 20GB of caching, so i've got 52GB of read cache on a 64GB system since the games and OS use less. It's kinda silly, but *also the reason i wanted the RAM in the first place
(he windows caching doesnt remember from a previous boot, and yeets itself if something needs it - while primocache hoards it like a dragon with RAM sticks for the shiny gold contacts

(*That, and the sale when i bought it. $400 Au for 2x32GB right before covid janked prices up)

Interesting SSD combo, i'm surprised that the QVO wasn't enough for the game - it's not like their read speeds are terrible.

PC's only been on 30 minutes and edge/TPU exclusively, but look at this nuts write reduction
If i had a QVO as an OS drive, i'd be running a long-ass delayed write on it.

Mussels, I did some perf. comparisons of the SATA 870 QVO 4 TiB and the sk-hynix P31 gold M.2 NVME SSD:

i. AnandTech Storage Bench - The Destroyer

ix. the average latency of the 870 QVO is 12 to 13 times slower than what the sk-hynix P31 gold is
x. the average read latency is 6 times slower than what the sk-hynix P31 gold provides
xi. the average write latency is 18 times slower than what the sk-hynix P31 gold provides
xii. 99th percentile latency is 14 times slower than what the sk-hynix P31 gold provides
xiii. the 99th percentile read latency is 8 to 9 times slower than what the sk-hynix P31 gold provides
xiv. the 99th percentile write latency is 3 times slower than what the sk-hynix P31 gold provides

j. AnandTech Storage Bench - Heavy

i. sk hynix gold P31 1 TB is 3x's faster than the 870 QVO in data rate when both SSD's are empty
ii. sk hynix gold P31 1 TB is 4 to 5x's faster than the 870 QVO in data rate when both SSD's are full

Every time the Samsung 870 QVO arbitrarily and without any notice disconnects itself from my system my primocache L2 cache gets wiped and I've noticed stuttering in Metro:Exodus and GTAV which isn't there once they're cached in the primocache L2 cache located on the sk-hynix P31 gold.

I wish I could try deferred writes with an L1 cache but I don't trust my memory overclocks and I don't have a UPS system.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
I mean yeah, i knew the QVO was slower - NVME vs SATA guarantees that as much as the QLC does.
But... ouch. That's pretty poor.

Every time the Samsung 870 QVO arbitrarily and without any notice disconnects itself from my system
This is more likely related to the stutter than anything else - faulty SATA cables or an issue with the SATA controller is most likely the cause there
If thats a ryzen system, that's just screaming SoC dropouts from errors on the PCI-E or IF buses

If you wanted to PM your system specs (motherboard and what sata port/controller the QVO is connected to) i can give you a quick rundown on some possibilities there, i've ran into a few oddities on intel and AMD platforms that may help it recover quicker, or avoid dropouts at all without clogging up this thread
 
Joined
Dec 12, 2020
Messages
1,755 (1.37/day)
I've had HDD's, SSD's and a DVD-RW intermittently disappear from device mangler and the BIOS with this z390 Taichi. I've tried adding dielectric grease to all the SATA connectors. Right now, I'm leaning more towards insufficient voltage to the PCH, but that's just a hope and a prayer. Maybe I should put everything on the Asmedia SATA controller! :)
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Joined
Feb 1, 2019
Messages
2,765 (1.41/day)
Location
UK, Leicester
System Name Main PC
Processor 13700k
Motherboard Asrock Z690 Steel Legend D4 - Bios 13.02
Cooling Noctua NH-D15S
Memory 32 Gig 3200CL14
Video Card(s) 3080 RTX FE 10G
Storage 1TB 980 PRO (OS, games), 2TB SN850X (games), 2TB DC P4600 (work), 2x 3TB WD Red, 2x 4TB WD Red
Display(s) LG 27GL850
Case Fractal Define R4
Audio Device(s) Asus Xonar D2X
Power Supply Antec HCG 750 Gold
Software Windows 10 21H2 LTSC
Installed on laptop now trial of the non pro version as not using L2 cache.

Added write only mode 512MB for C:, so caching browser, window log writes etc.

With browser not open sitting at desktop, it wrote 5.4MB of data to the cache simply from generic windows journal and log writes within the first minute, when it got flushed to disk it wrote 85% of it, and as I am typing this post that average is now 79% so circa 20% is rewrites which are saved via the write-defer cache.

Write-defer is in native mode, enabled flush on sleep, and defer latency is set to 60s for now. 4k cluster size.

Very nice to not see constant writes on the SSD in the task manager now.

Total write to disk just before i hit submit is now down to 43% cool, will update this thread later just before the laptop is put to sleep later tonight.
 
Joined
Dec 12, 2020
Messages
1,755 (1.37/day)
@Mussels and chrcoluk
Do you both have UPS's for your systems? I don't trust my memory overclocks enough to do deferred writes and I have already seen what marginal memory and a write L1 cache can do with Metro:Exodus: it corrupts savegames.
 
Joined
Feb 1, 2019
Messages
2,765 (1.41/day)
Location
UK, Leicester
System Name Main PC
Processor 13700k
Motherboard Asrock Z690 Steel Legend D4 - Bios 13.02
Cooling Noctua NH-D15S
Memory 32 Gig 3200CL14
Video Card(s) 3080 RTX FE 10G
Storage 1TB 980 PRO (OS, games), 2TB SN850X (games), 2TB DC P4600 (work), 2x 3TB WD Red, 2x 4TB WD Red
Display(s) LG 27GL850
Case Fractal Define R4
Audio Device(s) Asus Xonar D2X
Power Supply Antec HCG 750 Gold
Software Windows 10 21H2 LTSC
Laptop has built in battery and is backed up, PC has UPS and automatic hibernate when battery goes below target %, but currently only running this on my laptop. Both are running ram at its spec'd clocks.

If you dont trust your memory clocks, then dont keep using those clocks I guess. If they cause corruption with a L1 cache, then they going to cause corruption sooner or later without a L1 cache as well as the data still has to go through ram before writing.

Current stats for laptop this has mostly been idle, I have just loaded a couple of pages on prima games website for map guides for zestiria. Also loaded a google sheets page, browser has about 20 background tabs.

Plus a notepad open.

Total read 1.73GB
Total write 2.69GB (I thought consumer use was mostly reads eh?)
Total writes to disk 1.97GB (73.3%)
Unused cache 220MB so about 300MB utilised.

Interestingly even though its in write only mode for cache, it is reporting some read cache data.

115MB cached read (6.5%).
Unused and available cache read is of course 0

I assume these are reads taken from the write cache before its flushed.

So not as good saved writes as mussels.
 
Joined
Dec 12, 2020
Messages
1,755 (1.37/day)
@chrcoluk

I had never thought about it before but it was a good thing Metro:Exodus had problems with my memory overclocks+primocache L1 cache -- because my memory overclocks had passed 13 hrs. of memtest, 2.5 hrs. of prime95 large FFT's, 4 hrs. of 6 core orthos (also large FFT's) and windows 10's own memory test. It turned out all I had to do was increase the V-DIMM to 1.475V.

That is interesting that you saved on reads with a write only L1 cache! My L1 cache is set to R/W, but it's only 1 GiB.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Why not? Dielectric grease is perfect for keeping electrical connections free of corrosion and protect a connector. This is why it's been used in the automotive industry for decades.
bceause it's not designed for the metals used in PC's and you get reactions like what happened with the EK/asus VRM catastrophe

It's a non conductive lubricant, and PC hardware DOESNT NEED LUBRICANTS. The secondary use to prevent corrosion doesnt apply when PC components don't use the same metals as those industries - worst case it's going to add a sticky layer to connect contaminants only to end up jammed inside the opposing connector

Why aren't you filling the charging port on your phone with vaseline, your wall sockets for mains power, or coating your fusebox?
Because it's not meant for those situations.


It's also entirely offtopic for primocache, other than that one guy who does this wondering why he has issues - that are likely because he's screwed his devices with the grease.
 
Joined
Jul 5, 2013
Messages
25,764 (6.45/day)
bceause it's not designed for the metals used in PC's and you get reactions like what happened with the EK/asus VRM catastrophe
What? Dielectric grease is chemically inert. It can't react with metals unless exposed to heavy alkali materials, but then you'd have a much bigger problem.
It's a non conductive lubricant, and PC hardware DOESNT NEED LUBRICANTS.
You mean fan bearings aside? However, there is nothing wrong with protecting contacts from the elements, which dielectric grease does by design.
with vaseline
Vaseline is VERY different from dielectric grease. The chemical properties and formulations are not comparable.

It's also entirely offtopic for primocache
You're right we're getting off topic..
 
Joined
Mar 21, 2021
Messages
4,587 (3.89/day)
Location
Colorado, U.S.A.
System Name CyberPowerPC ET8070
Processor Intel Core i5-10400F
Motherboard Gigabyte B460M DS3H AC-Y1
Memory 2 x Crucial Ballistix 8GB DDR4-3000
Video Card(s) MSI Nvidia GeForce GTX 1660 Super
Storage 500GB SN550 WD Blue NVMe M.2
Display(s) Dell P2416D (2560 x 1440)
Power Supply Apevia 600W Gold 80 Plus
Software Windows 11 Home
Not wanting to get off topic, but felt it should be said.
  • Dielectrc grease is silicone based and so unlike vaseline it will not attack carbon based plastics.
I actually use silicone oil for PC use.

Now back to the topic: rather than buy Primocache I would use the $30 saved toward a bigger solid state drive.
 
Joined
Aug 30, 2006
Messages
7,200 (1.11/day)
System Name ICE-QUAD // ICE-CRUNCH
Processor Q6600 // 2x Xeon 5472
Memory 2GB DDR // 8GB FB-DIMM
Video Card(s) HD3850-AGP // FireGL 3400
Display(s) 2 x Samsung 204Ts = 3200x1200
Audio Device(s) Audigy 2
Software Windows Server 2003 R2 as a Workstation now migrated to W10 with regrets.
Now back to the topic: rather than buy Primocache I would use the $30 saved toward a bigger solid state drive.
Spend the $30 on a small nvme drive, point your cache and pagefile and tmps there. Save ssd thrashing. Have a faster pagefile than leaving on ssd. For additional speedup, if you have loads of ram, allocate some to a ramdisk. There are free versions available.
 
Joined
Feb 20, 2020
Messages
9,340 (5.93/day)
Location
Louisiana
System Name Ghetto Rigs z490|x99|Acer 17 Nitro 7840hs/ 5600c40-2x16/ 4060/ 1tb acer stock m.2/ 4tb sn850x
Processor 10900k w/Optimus Foundation | 5930k w/Black Noctua D15
Motherboard z490 Maximus XII Apex | x99 Sabertooth
Cooling oCool D5 res-combo/280 GTX/ Optimus Foundation/ gpu water block | Blk D15
Memory Trident-Z Royal 4000c16 2x16gb | Trident-Z 3200c14 4x8gb
Video Card(s) Titan Xp-water | evga 980ti gaming-w/ air
Storage 970evo+500gb & sn850x 4tb | 860 pro 256gb | Acer m.2 1tb/ sn850x 4tb| Many2.5" sata's ssd 3.5hdd's
Display(s) 1-AOC G2460PG 24"G-Sync 144Hz/ 2nd 1-ASUS VG248QE 24"/ 3rd LG 43" series
Case D450 | Cherry Entertainment center on Test bench
Audio Device(s) Built in Realtek x2 with 2-Insignia 2.0 sound bars & 1-LG sound bar
Power Supply EVGA 1000P2 with APC AX1500 | 850P2 with CyberPower-GX1325U
Mouse Redragon 901 Perdition x3
Keyboard G710+x3
Software Win-7 pro x3 and win-10 & 11pro x3
Benchmark Scores Are in the benchmark section
Spend the $30 on a small nvme drive, point your cache and pagefile and tmps there. Save ssd thrashing. Have a faster pagefile than leaving on ssd. For additional speedup, if you have loads of ram, allocate some to a ramdisk. There are free versions available.
Hi,
Small hell some 1tb are going for 60.us lol
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Now back to the topic: rather than buy Primocache I would use the $30 saved toward a bigger solid state drive.
Or buy primocache and have those SSD's live twice as long - or as others said, use a spare drive to cache the whole system.

I've got spare 512GB NVME drives here that with a write buffer would make for one hell of a NAS acclerator, if i wasnt limited by being on gigabit in the first place making it pointless.
 

S2G-Unit

New Member
Joined
Apr 17, 2023
Messages
7 (0.02/day)
Advice needed for my setup. I've read all 6 pages & decided to order a 118GB P1600x. All I care about is the lowest latency & fastest Random 4k reads on my system. (I have no use for large file transfers/bandwidth)

Ryzen 7700x
Samsung 990 Pro
6TB HDD
Gigabyte B650 Aorus Pro
32GB DDR5 @ 6200 CL30

For the lowest latency & fastest Random 4k reads in Windows just for overall system quickness when using programs work/business (More for fun of a faster system. I have no use for large file transfers/bandwidth. I did notice and really did enjoy the upgrade from a 860 EVO to the 990 Pro).

Should I:
1. Put my OS on the p1600x?
2. Keep my OS on the 990 Pro & use the P1600x with PrimoCache?
3. If i don't mind spending the money, get a 905p from newegg for 400$ USD? Any real advantage to using a 905p for my OS & cache vs just having a p1600x for cache? I think im wondering what would be the point in getting a 960gb Optane drive if I can use PrimoCache?
4. Use options 1-3 then dedicate maybe 6-12GB of ram as a ramdisk?

What are the benefits & downsides to each?

Was also curious if anybody has benchmarked & compared random 4k read speeds on the same system with a P1600x & a 900\905p?
How much faster are the 900p/905p than P1600x in random 4k reads?
 
Last edited:
Joined
Apr 18, 2019
Messages
2,172 (1.15/day)
Location
Olympia, WA
System Name Sleepy Painter
Processor AMD Ryzen 5 3600
Motherboard Asus TuF Gaming X570-PLUS/WIFI
Cooling FSP Windale 6 - Passive
Memory 2x16GB F4-3600C16-16GVKC @ 16-19-21-36-58-1T
Video Card(s) MSI RX580 8GB
Storage 2x Samsung PM963 960GB nVME RAID0, Crucial BX500 1TB SATA, WD Blue 3D 2TB SATA
Display(s) Microboard 32" Curved 1080P 144hz VA w/ Freesync
Case NZXT Gamma Classic Black
Audio Device(s) Asus Xonar D1
Power Supply Rosewill 1KW on 240V@60hz
Mouse Logitech MX518 Legend
Keyboard Red Dragon K552
Software Windows 10 Enterprise 2019 LTSC 1809 17763.1757
Advice needed for my setup. I've read all 6 pages & decided to order a 118GB P1600x. All I care about is the lowest latency & fastest Random 4k reads on my system.

Ryzen 7700x
Samsung 990 Pro
6TB HDD
Gigabyte B650 Aorus Pro

For the lowest latency & fastest Random 4k reads in Windows.
Should I:
1. Put my OS on the p1600x?
2. Keep my OS on the 990 Pro & use the P1600x with PrimoCache?

What are the benefits & downsides to each?
OS on the 990pro and the P1600X as cache will extend the life of the 990pro considerably.

Theoretically, it should also be the 'overall fastest'.
Caching transactions will happen very quickly off the P1600X, and transactions exceeding/not in the cache would benefit from the 990pro's bandwidth.

[Wear does seem to be an issue (even after the firmware updates) on the 980pro and 990pro both.]
 
Joined
Mar 21, 2021
Messages
4,587 (3.89/day)
Location
Colorado, U.S.A.
System Name CyberPowerPC ET8070
Processor Intel Core i5-10400F
Motherboard Gigabyte B460M DS3H AC-Y1
Memory 2 x Crucial Ballistix 8GB DDR4-3000
Video Card(s) MSI Nvidia GeForce GTX 1660 Super
Storage 500GB SN550 WD Blue NVMe M.2
Display(s) Dell P2416D (2560 x 1440)
Power Supply Apevia 600W Gold 80 Plus
Software Windows 11 Home
How much CPU power does Primocache take up?
 
Joined
Jul 5, 2013
Messages
25,764 (6.45/day)
2. Keep my OS on the 990 Pro & use the P1600x with PrimoCache?
This would likely work best, however @Mussels would likely be the best person to ask as he seems to have the most experience with this program. I personally would defer to his input and advice.

How much CPU power does Primocache take up?
Not much. As far as CPU resources go, it's very light-weight.
 
Last edited:

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.12/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
Advice needed for my setup. I've read all 6 pages & decided to order a 118GB P1600x. All I care about is the lowest latency & fastest Random 4k reads on my system. (I have no use for large file transfers/bandwidth)

Ryzen 7700x
Samsung 990 Pro
6TB HDD
Gigabyte B650 Aorus Pro
32GB DDR5 @ 6200 CL30

For the lowest latency & fastest Random 4k reads in Windows (I have no use for large file transfers/bandwidth).

Should I:
1. Put my OS on the p1600x?
2. Keep my OS on the 990 Pro & use the P1600x with PrimoCache?
3. If i don't mind spending the money, get a 905p from newegg for 400$ USD? Any real advantage to using a 905p for my OS & cache vs just having a p1600x for cache? I think im wondering what would be the point in getting a 960gb Optane drive if I can use PrimoCache?
4. Use options 1-3 then dedicate maybe 6-12GB of ram as a ramdisk?

What are the benefits & downsides to each?

Was also curious if anybody has benchmarked & compared random 4k read speeds on the same system with a P1600x & a 900\905p?
How much faster are the 900p/905p than P1600x in random 4k reads?
This isn't an SSD benchmarking thread, so not a lot of people are going to be here with that sort of specific hardware knowledge
primocache can be used for different layers, with RAM first, then you go down in order of speed - so an NVME drive can cache a RAID array of mechanicals


My best luck has been the OS on it's own drive with a write-only cache to reduce wasted writes thanks to it being TRIM aware - i'm at ~55% writes saved today
1681811555332.png


Games drive has a 75% hit on reads, with just 32GB for a 2TB games drive.
1681813705404.png



It just comes down to scale and what you're trying to achieve - for reduced writes and extended lifespan, you want a long RAM write cache. For sped up reads, you need a low latency cache as big as possible (RAM or high speed NVME) caching whatever is downstream

How much CPU power does Primocache take up?
Hard to say, for my usage it's extremely low. The bigger the block size you use, the lower the RAM overheads and lower the CPU usage.
Running benchmarks for example can be misleading, because you'll see higher CPU usage since the storage part of the test is no longer a bottleneck

Lower block size tends to result in faster file system performance (since it loads/moves/writes exactly what it needs, and no more)

Caching big writes will result in removing many other files from the cache, so lock set %'s in.

How much it helps comes down to repetition - windows (and browsers) writes a lot of small crap, so a cache helps reduce the waste.
Reads? a 128GB cache wont do much for a collection of movies, but it might help a games drive by caching all the exes, dlls and other small files that 4k random read speeds are normally poor at.
 
Joined
Aug 30, 2006
Messages
7,200 (1.11/day)
System Name ICE-QUAD // ICE-CRUNCH
Processor Q6600 // 2x Xeon 5472
Memory 2GB DDR // 8GB FB-DIMM
Video Card(s) HD3850-AGP // FireGL 3400
Display(s) 2 x Samsung 204Ts = 3200x1200
Audio Device(s) Audigy 2
Software Windows Server 2003 R2 as a Workstation now migrated to W10 with regrets.
Should I:
1. ..
2. ..
3. ..
4. ..
5. Upgrade to 64GB system RAM

Exactly where do you "feel" the need for speed on your PC? General use, or specific situations. A ramdisk for internet temps makes any browsing experience very snappy, and saves SSD thrashing. Booting into windows 0,5 second faster is a once-in-a-rare time event, whereas what you are doing all day makes a difference.
 
Status
Not open for further replies.
Top