Category Archives: Benchmarking

Thinking About Windows 10/11 SSDs

I’m still busy benchmarking away on the two Thunderbolt4/USB4 PCs that Lenovo has recently sent my way. But as I’ve been doing so, I’ve been thinking about Windows 10/11 SSDs in general. On that path, I’ve realized certain principles that I’d like to share with you, dear readers.

I’m spurred in part to these statements from a sponsored (and pretty contrived) story from MSPowerUser entitled “Is NVMe a Good Choice for Gamers?” My instant response, without reading the story — which actually focuses on storage media beyond the boot/system drive — was “Yes, as much as you can afford.” Spoiler alert: that’s what the story says, too.

Where Thinking About Windows 10/11 SSDs Leads….

Here are some storage media principles that flow from making the most of a new PC investment.

  1. The more you spend on a PC, the more worthwhile it is to also spend more on NVMe storage.
  2. Right now, PCIe Gen4 drives run about 2X the speed of PCIe Gen3 drives. They don’t cost quite twice as much. Simple economics says: buy the fastest NVMe technology your PC will support.
  3. Buy as much NVMe storage as you can afford (or force yourself to spend). For pre-built PCs and laptops, you may want to buy NVMe on the aftermarket, rather than get the drives pre-installed. Markup on NVMe drives can be painful. Hint: I use Tom’s Hardware to keep up with price/performance info on NVMe SSDs and other PC components (it’s also the source for the lead-in graphic for this story, which still prominently displays the now-passe Intel Optane as an SSD option. Caveat emptor!).
  4. Corollary to the preceding point: fill every M.2 slot you can in your build. For both my recent Lenovo loaners — the P360 Ultra and the P16 Mobile Workstation — that means populating both slots with up to 4TB each. Right now, the Kingston KC3000 looks like a 4TB best buy of sorts.

Thinking Further (and Outside the Box)

More thoughts in this vein, with an eye toward external drives and multi-tiered storage (archives and extra backups):

  1. If you’re going to put an NVMe SSD in an external enclosure, you will be OK for the time being in a USB 3.2 rather than a USB 4 enclosure. Right now, the newer enclosures cost more than twice as much but don’t deliver anywhere near 2x the speed (except on synthetic benchmarks — I used C: imaging times as a more reliable indicator). Over time this will no doubt change, and I’ll keep an eye on that, too.
  2. I don’t consider spinners (conventional mechanical hard disk drives, or HDDs) any more, except for archival and inactive storage. If I need something for work or play, it goes on an SSD. If I might need something, someday (or to restore same) then it’s ok on an HDD.

I used to restrain spending on NVMe SSDs because of its high price differential. I’m now inclined to believe that restraint is a false economy and forces less productivity as a result. That’s why I’m rethinking my philosophy. I haven’t quite yet gotten to Les Blanc’s famous dictum (“Spend It All”) but I am coming around to “Spend As Much as You Can”…

Remember This Fundamental Assumption, Tho…

My reasoning aims at high-end PCs where users run data-, graphics-, and/or compute-intensive workloads. It does not apply, therefore, to home, hobbyist, and low-end office users. For them typical productivity apps  (e.g. MS Office or equivalent), email, web browsing and so forth predominate. They wouldn’t need, nor benefit much from, buying lots of fast NVMe storage. That said, a 1 TB fast-as-possible NVMe for the boot/system drive is the baseline. Other storage options will balance themselves against budget to dictate other choices and PC builds for such users.

In different terms, if you’re not maxing out your PC running data analytics, 3D models and other high-end graphics rendering, or AI or machine learning stuff, this advice is most likely overkill. Too, too costly. But for this user community, more spent on NVMe (and GPUs and memory as well) will repay itself with increased productivity. ‘Nuff said.

Facebooklinkedin
Facebooklinkedin

Backblaze Data Confirms SSD Trumps HDD Reliability

It’s always made sense on an intuitive basis. Hard Disk Drives (HDDs) include spinning platters, moving arms with read/write heads, motors to power things, and gears to control action. SDDs are made entirely of circuitry: no moving parts. Thus, it’s compelling to assert that SDDs should be more reliable, and less prone to failure than HDDS. And indeed, the latest 2022 Drive State report from online backup and storage provider Backblaze weighs in on this topic. As I read it, that Backblaze data confirms SSD trumps HDD reliability.

The lead-in graphic shows 4 years’ worth of SSD data vs. 8 years for HDDs for boot drivers in their thousands of datacenter based servers. Whereas there’s a dramatic upward knee in the curve for HDDS between years 4 and 5 (from 1.83% to 3.55%), failures actually dipped for SDDs during that interval (from 1.05% to 0.95%). Interesting!

How Backblaze Data Confirms SSD Trumps HDD Reliability

The afore-linked report explains that boot drives function in multiple roles on the company’s plethora of storage servers. They store log and temprorary files; they maintain storage holdings based on each day’s storage activities and volume. The disparity in the number of years for which data is available comes from later adoption of SDDs as boot drives at BackBlaze. That practice started in Q4 2018. Today, all new servers boot from SSDs; older servers whose HDD boot drives fail get SSD replacements.

The numbers of SSDs keep going up, too. The end-of-year 2021 SSD report encompassed 2,200 SSDs. By June 30, 2022, that count grew to 2,558. Failure rates for such devices show much lower numbers than for HDD (see the tables labeled Backblaze SSD Quarterly Failure Rates in the latest report for more detail). Models included come from the following vendors: Crucial, Dell, Micron, Seagate and WDC.

Note: the report itself says:

For any given drive model in this cohort of SSDs, we like to see at least 100 drives and 10,000 drive-days in a given quarter as a minimum before we begin to consider the calculated AFR to be “reasonable”.

The real news, of course, is that quarterly, annualized and lifetime failure rates for SSDs are significantly lower than for HDDs, based on Backblaze’s own long-running data collection. Thus their conclusion comes with the weight of evidence “…we can reasonably claim that SSDs are more reliable than HDDs, at least when used as boot drives in our environment.”

Good stuff! As for me, I like SSDs not just because they’re less prone to failure. They’re also FAST, if more expensive per storage unit than spinners.

 

 

Facebooklinkedin
Facebooklinkedin

Thunderbolt Dock Loses GbE Port

Drat! In jacking around with my Belkin Thunderbolt 3 Dock Plus today, I couldn’t help but notice that the wired Ethernet port wasn’t blinking. Further testing included multiple cables and connections to the same port, none of which worked. When I tried a passive Thunderbolt 3 mini-dock in the other USB-C port on the Lenovo X12, that wired Ethernet port worked immediately. Thus, I can only conclude that Thunderbolt Dock loses GbE port is the right diagnosis. Sigh.

Note: The lead-in graphic for this story shows the rear view of the aforementioned Belkin device, with its RJ-45/GbE port at the left. No blinkin’ lights, man!

If Thunderbolt Dock Loses GbE Port, Then What?

For the time being, I’m using another dock — the Thunderbolt 3 Minidock — just for its RJ-45 GbE connection. Good thing my X12 Hybrid has a spare USB-C/Thunderbolt port, eh?

Longer term, I’ve already contacted Belkin about sending me a replacement. They’ve got a nice looking Thunderbolt 4 dock for sale now, so hopefully they’ll ship one my way. I’ve also gone ahead and ordered the CalDigit TS4, reputedly one of the best Thunderbolt 4 docks on the market today.

Thunderbolt 4 Docking Brings Other Benefits

Acquiring one or more Thunderbolt 4 docks will also help with my ongoing testing of NVMe SSD enclosures. As I reported a few days ago, switching from USB-C/3.1 or 3.2 to Thunderbolt 3 makes a difference in IO performance on my fastest SSD enclosure/drive combos. I’m curious to see if a bump to Thunderbolt 4 will make any additional difference.

According to what I read, throughput doesn’t vary that much for external drives from Thunderbolt 3 to 4. I’ve also observed that synthetic IO tests (e.g. CystalDiskMark) tend to overstate the real-world speed-ups available from faster buses. Thus it will be interesting to observe exactly how much difference the bump from 3 to 4 makes.

Stay tuned! I’ll let you know what comes of that testing. Should be fun!

Facebooklinkedin
Facebooklinkedin

Samsung NVMe Drive Failing

In a recent story here, I mentioned a possible mismatch between some components. On the one hand: an old Samsung MZVPV512HDGL OEM NVMe drive. On the other hand: a brand-new PCIe x4 USB 3.2/Thunderbolt NVMe enclosure. Upon swapping in a newer ADATA drive my issues with the enclosure vanished. So I mounted the other drive in an older Sabrent NVMe enclosure. Now I’m getting indications of the Samsung NVMe drive failing. A strong indicator shows up as the lead-in graphic above.

What Says: Samsung NVMe Drive Failing?

The inability to perform write tests using HD Tune is a pretty big tell. Interestingly, though: chkdsk and CrystalDiskInfo both report the drive as healthy. My best guess is that write failures are occurring, and that HD Tune won’t “write past” such things, while the other tools rely on SMART data and surface analysis and aren’t seeing active errors.

My plan is to retire the drive as soon as the replacement part shows up. That’s been en route via Amazon for too long now, so I just cancelled that order and placed a new one. Hopefully it will be here tomorrow, including a 1TB Sabrent Rocket 4 Plus with internal read/write speeds of up to 6+/4+ Gbps. Of course, that’s not gonna happen in a USB 3.2/Thunderbolt enclosure. But I am darn curious to see how fast the bus can go when the drive is fast enough to get out of the way.

Stay Tuned: More to Come!

According to what I read online, I may be able to get read/write speeds in excess of 2 Gbps via Thunderbolt 3 from the NVMe enclosure. So far, the best I’ve seen from my older Sabrent (USB 3.2 only) enclosures is on the order of 1.1 Gbps. So it should be pretty easy to tell if the new drive/enclosure speeds things up.

Facebooklinkedin
Facebooklinkedin

Thunderbolt 3 SSD Enclosure Raises Odd Issues

This is the part of playing with Windows that I love best. I’m researching different speeds for backup drives, ranging from a USB-C HDD drive caddy into the SSD realm. My objective is to see how fast an external USB-C drive can go, and to see if Thunderbolt support makes any difference.  I added a cheapo (US$29) NVMe enclosure to my line-up. But alas, that Thunderbolt 3 SSD enclosure raises odd issues. Let me explain…

Why Thunderbolt 3 SSD Enclosure Raises Odd Issues

As far as I can tell, I went too far back in time with the first M.2 NVMe I tried out in the cheapo new NVMe enclosure. My initial attempt featured a 2016 vintage OEM Samsung MZVPV512HDGL SSD. It kept blowing up during write testing in CrystalDiskMark, and it wouldn’t make a Macrium Reflect backup.

So I cannibalized a newer ADATA XPG 256GB SSD (vintage 2020) from my Sabrent-enclosed Ventoy drive and tried that instead. It worked just fine, and got aggregate read/write speeds from Reflect of 5.7/3.0 Gbps when backing up my Lenovo X12 Hybrid Tablet. It includes a USB 3.2 version of USB-C with Thunderbolt 3 support. Total backup time on that system was 6:16 with 76.5 GB on the C: drive and under 1.5 GB on the other partitions. Figure 78 GB overall, that produces a physical time (no compression) of roughly 200 Mbps of ongoing read/write activity. By comparison an mSATA drive (vintage 2013) takes just under16 (15:56) minutes to complete the same backup. That’s more than 60% faster!

It’s All About the Speed

My best guess is that the older drive wasn’t sufficiently compatible with the PCIe x3/x4 requirements inside the NVMe enclosure. Once I switched over to something newer (and definitely PCIe x3 compliant), everything worked fine. I’ve got a brand-new PCIe x4 SSD coming today or tomorrow, and am hopeful the faster media will also produce faster transfer rates for backup, too. We’ll see!

Facebooklinkedin
Facebooklinkedin

Why USB Disk Speeds Matter

It’s been a busy and interesting week. I’ve been messing around with numerous backups and restores. Ditto for mounting ISOs and running Windows repair installs. A LOT of disk reads and writes to USB drives have been involved. Because of the huge amounts of data involved, I’m better prepared to explain why USB disk speeds matter. A LOT!

Why USB Disk Speeds Matter So Very Much

In a word, the shortest possible answer is “Time.” If you can get something done faster, you can do more in a single work interval. Compare the USB disk speeds for an NVMe drive in a USB-C enclosure (left) to those for an mSATA drive in a USB-A 3.1 enclosure (right — see lead-in graphic). When backups and restores are concerned the top lines (which involve large file transfers) actually matter. Of course, all the times matter as well.

But those differences are pretty stark for backup and restore. Let me explain… If you look at the top pairs of numbers, these cover large data transfers with a queue depth of 8 (upper) and 1 (lower). In both pairs of numbers, the NVMe drive is over twice as fast as the mSATA drive. Those same results were born out in backups and restores (7 and 14 minutes for backup; 11 and 23 minutes for restore).

The More You Do, the Better You’ll Like It!

Those results show why I’ve long been a believer in using fast USB drives whenever possible. I’m still waiting to see what kind of bump I can get with a Thunderbolt 4 NVMe enclosure, proper cables and enclosure, and Thunderbolt 4 on the host device. From what I read, it should be 25-40% as fast again.

This realization came to me when I started copying a backup from a BitLocker protected NVMe drive to an mSATA unprotected drive. I got a consistent 26-27 MBps transfer rate between the two devices. It took over 20 minutes to copy the file!

If I could’ve gone Thunderbolt 4 all the way, I could have quadrupled the transfer speed or better. That would cut my wait time from 20 minutes to 5. Waiting for necessary data can’t be completely bypassed — but it surely shows the “need for speed” on such occasions.

Facebooklinkedin
Facebooklinkedin

Windows 11 Sports Slow NVMe Driver?

Here’s an interesting thread emerging from the Windows press. A growing number of outlets are reporting that Microsoft’s own NVMe drivers run slower than their Win10 counterparts on identical (and other hardware). If Windows 11 sports slow NVMe driver, what can users do? Not much, it turns out, unless they can run a third-party driver instead (e.g. Samsung NVMe Controller). For good coverage on this topic, see Taras Buria’s recent WinAero story “Windows 11 apparently slows down NVMe SSDs.” It cites a range of interesting and informative original sources.

If Windows 11 Sports Slow NVMe Driver, Then What?

What appears to be affected is the OS boot/system drive (usually C:, where Windows itself resides). Some independent tests show that other non-OS partitions don’t suffer performance degradation. But OS partitions could suffer from reductions in random read/write speeds of 50% or worse. For grins I compared CrystalDiskMark stats from my 11th gen Lenovo X12 Hybrid Tablet running Windows 11 to my 6th gen home-brew Z170-based desktop. The former has a WD SN530 1 TB SSD, while the latter has a Samsung 950 1 TB SSD.

As you can see in the lead-in graphic, the newer Windows 11 unit is a bit slower on most readings than the older Windows 10 PC. Indeed QD32 random reads  are about 1/3 slower. That said, random writes of the same ilk go the other way (but with a less-than-7% delta). For random reads/writes with QD1, 11 edges 10 on writes by just over 9%, and vice-versa for reads by just over 15%. Kind of a wash, if you ask me.

What This Means for Upgrade Plans

MS has acknowledged that the issue is known to them and that they’re working on a fix, ETA unknown. Some reports aver that this phenomenon justifies postponing upgrades until a fix is in. My own experience with Windows 11 has been uniformly positive so far, NVMe performance observations notwithstanding. I’d recommend rethinking upgrades on PCs with heavy I/O workloads (e.g. CAD, AI, data analysis, and so forth). But for routine personal or productivity computing, it doesn’t really seem to make a noticeable difference.

I’ll be watching this issue as it unfolds. Count on me to let you know when this situation changes. Given the importance of NVMe to modern computing workloads, lots of people will no doubt follow this carefully and closely.

Note Added December 10: KB5007262 Fixes Issue, But…

KB5007262 should be installed as part of the upcoming December 14 updates for production Windows 11. It came out for Beta/Dev Channel users in November (ditto for production versions, as a Preview Update on November 23). Indeed, it seems to fix the performance issues. According to this WinAero story, the MS bugfix info for the KB5007262 announcement includes the following text:

“Addresses an issue that affects the performance of all disks (NVMe, SSD, hardisk) on Windows 11 by performing unnecessary actions each time a write operation occurs. This issue occurs only when the NTFS USN journal is enabled. Note, the USN journal is always enabled on the C: disk.”

And indeed, on my X1 Extreme laptop (8th gen i7, Samsung OEM 512GB NVMe SSD, 32 GB RAM) speeds are where experience teaches me they should be. According to the afore-linked story, this fix applies to all versions of Windows 11 except Dev Channel. As far as I can tell, that version remains subject to the slow-down. I’m looking for some additional word from MS on this topic. Hopefully, they’ll fix it there soon, too. Stay tuned!

Facebooklinkedin
Facebooklinkedin

Estimating Windows 11 Restart Time

Because today is Patch Tuesday (November 9) I got several opportunities to see WU at work handling updates. Owing to across-the-board Cumulative Updates (CUs) today, that meant 1 production version, 1 Beta version and 2 Dev Channel versions. As you can see from the lead-in graphic, estimating Windows 11 restart time is now part of what WU offers. I thought this was pretty cool, until I realized all 4 PCs proffered the same estimate.

What Does Estimating Windows 11 Restart Time Tell You?

Just for grins, I timed a couple of my restarts to see how long they would actually take. I’m pleased to report that the MS/Win11 estimate is conservative. It took 1:25 to get to the desktop with GadgetPack running to show me a second hand on its clock widget on my X1 Extreme (i7-8850H CPU, 6 Cores). It took 2:35 to get to the same place on my X380 Yoga (i7-8650U CPU, 4 Cores).

That tells me that MS isn’t necessarily driving the estimate from observation of previous start times. Rather, it looks like a rough-and-ready interval that will not set user expectations overly high. Why do I say these things? Because the number was the same across a range of CPUs. And because the number was too high for all of them.

My gut feel is that if this estimate were data driven, it would be slightly high on some and slightly low on others. Because it was the same for all four PCs, and too pessimistic likewise, it strikes me as a “safe estimate” probably based on worst-case observations.

How Does 4 Minutes Strike YOU?

All this said, I think 4 minutes is neither a terrible number nor a glorious one. When I’ve really worked at getting start-up times to their barest minimums on Windows 10 (haven’t yet tried this on 11) I’ve seldom gotten below 1:30 or so. But I’ve read about others who’ve documented start times just under a minute (0:45 or higher).

This may be a fruitful topic for research and play. Now, I just need to find the necessary spare time…

Facebooklinkedin
Facebooklinkedin

Win10 Rollback Works But Thunderbolt Issues Continue

Big Sigh. I’ve been trying to get the Thunderbolt 4 firmware updated on the snazzy new Lenovo ThinkPad X1 Carbon Gen 9 they sent me, but to no avail. Today, I observed that Win10 rollback works but Thunderbolt issues continue. Something gets weird when the PC reboots to do the firmware install. I see a short (and tiny) error message long enough to know it’s there, but definitely not long enough to read it, or interpret its significance.

When Win10 Rollback Works But Thunderbolt Issues Continue, Then What?

First, the good news. I elected to roll back my Windows 11 update on this machine and it not only went well, it finished in under 3 minutes. That’s amazing! It also confirmed that the Windows.old snapshot is of whatever vintage and state the OS was at the time of upgrade. All my account stuff remained clear and workable, thank goodness.

Now, the bad news. I remain unable to complete the firmware update successfully. That means Thunderbolt sees no devices on either of the PC’s two USB-C/Thunderbolt 4 ports. Bummer! It also means I’m sending this fish back to the pond (Lenovo, that is) with a request to return it when THEY can fix this driver issue. For me, Thunderbolt 4 is a big deal. I don’t think I can review this system without a working and capable Thunderbolt 4 connection for me to test performance, throughput, and so forth.

That said, the USB-3 Type A port is remarkably fast. I get better performance out of my old, tired mSATA drives on this machine (Samsung EVO SSDs in Sabrent mSATA enclosures) than I’ve ever seen before.

Do All Things Come to He Who Waits?

I guess I’ll be finding out. Tomorrow, I’ll fire off an email to the reviews coordinator, explain my situation, and let them know I’m sending the laptop back. It will be absolutely fascinating to see how they respond. I’m hopeful I’ll get a fixed (or replacement) laptop soon. If and when I do, I’ll start posting madly about what I see and learn. Right now, I just can’t go forward with a major subsystem on the fritz. Hope that makes sense…

Facebooklinkedin
Facebooklinkedin

Loaner Laptop Poses Weird USB Situation

I took delivery of a nifty new laptop here at Chez Tittel late last week. Among the zillions of other things going on around here, I’ve been fooling with this machine since it arrived. This loaner laptop poses weird USB situation, though: I get faster throughput from its USB-A 3.2 Gen1 port than either of its USB4 Type-C/Thunderbolt 4 ports. Throughput is about 10X faster on the USB-A port than on USB-C. That’s not how it’s supposed to work. Go figure!

Driver Issues Explain How Loaner Laptop Poses Weird USB Situation

Once I realized what was going on. I jumped into Device Manager. Sure enough there’s an issue with the ThinkPad Thunderbolt Retimer Firmware. Whaddya bet this could impact USB-C/Thunderbolt 4 timing?

And then, things get more interesting. Lenovo Vantage thinks the firmware update is already installed. Device Manager shows “Firmware update was unsuccessful.” Attempts to uninstall/reinstall don’t work, and manual installation of the downloaded firmware package N32TT02W.exe from Lenovo Support don’t work either.

I need some firmware juju. So I’m contacting Lenovo Support to see what they can tell me. I’ll admit I got fooled when Vantage told me the update was installed (and didn’t check DevMgr until later). Now, it looks like I’ll have to roll the machine back to Windows 10 so I can make sure the update gets properly applied. And then, I’ll roll forward again to Windows 11. Just another day in the life, here in Windows-World!

Checking Updates, Post Install

It hasn’t eluded me that checking the firmware install before upgrading to 11 would have been a peachy idea. I’m not one to rush into such things normally. But I wanted to see how the new PC would work with the new OS. I guess I’m  starting to understand there’s at least one good reason why Lenovo didn’t send me the device with Windows 11 already installed.

As I look around the Lenovo site, I see they have Thunderbolt drivers for Windows 11 aplenty. It’s just that they don’t have one for my X1 Carbon Gen 9 laptop just yet. Live and learn, dear readers: that’s why I’m going to try to do.

Facebooklinkedin
Facebooklinkedin