Thunderbolt 5 VS Thunderbolt 4 SSD

[@DavidHarry]

TB501 enclosure ➡ bit.ly/4fsUfEQ TBU405Pro M1 Enclosure ➡ bit.ly/4i1edsR Acasis on Amazon: TB501 enclosure ➡ Not available yet on Amazon TBU405Pro M1 Enclosure ➡ amzn.to/41V6p62 WD SSDs: SN850X 4TB ➡ geni.us/SN850X-4TB SN850X 2TB ➡ geni.us/SN850X-2TB SN850X 1TB ➡ geni.us/SN850X-1TB Samsung SSDs: 990 Pro 4TB ➡ geni.us/990PRO4TB 990 Pro 2TB ➡ geni.us/990PRO-2TB 990 Pro 1TB ➡ geni.us/990PRO1TB Links to Apple products: M4 Max MacBook Pro 16" ➡ geni.us/70Xn M4 Max MacBook Pro 14" ➡ geni.us/57QzTfu M4 Pro MacBook Pro 16" ➡ geni.us/qSYR M4 Pro MacBook Pro 14" ➡ geni.us/4thf yu0 M4 MacBook Pro 14" ➡ geni.us/Y7TFR Mac Mini M4 ➡ geni.us/oibaLH Mac Mini M4 Pro ➡ geni.us/ImJto Video chapters: 00:00 Intro and explanation 02:01 Thunderbolt 4 speed test 05:08 Thunderbolt 5 speed test 07:45 Speed test results 09:21 End summary Thunderbolt 5 SSDs and Thunderbolt 4 SSDs are both ways to get very fast external storage drives for any M4 Mac that has Thunderbolt 5 ports. Including the M4 Max MacBook Pro, which is what I use in this video and any M4 Pro Apple Silicon Mac such as the M4 Pro Mac. You can also use either a Thunderbolt 4 SSD or even a Thunderbolt 5 SSD with any Apple Silicon Mac that only has Thunderbolt 4 ports, such as any M1 Mac, M2 Mac, M3 Mac or base M4 Mac. However, using a Thunderbolt 5 SSD with any Thunderbolt 4 Mac will only give you Thunderbolt 4 speeds from a Thunderbolt 5 SSD. The results of these speed tests should hopefully give you a good idea as to the real world differences between a Thunderbolt 4 SSD and Thunderbolt 5 SSD when used as an external storage drive for your Mac. For anyone who has seen some of my recent speed test videos. I did not do the Blackmagic synthetic speed tests in this video as the video was specifically about real world storage usage. I also didn't show the formatting either, again, because this was a comparison video and not necessarily a build video. Plus I wanted to make a short video. Although, short is a relative term, as a 19 minute video is short for me but probably not most other KZbinrs 🤣 Anyway, please let me in the comments if you like the longer videos with more detail. I'm just not sure if they are too boring for most people. Thanks, Dave. I will be doing a long video soon showing how to put together the Acasis TB405Pro M1 Thunderbolt 4 to NVMe enclosure with the 4TB WD SN850X Gen 4 NVMe SSD. Basically, the one used in this video. In that video I will also go through the formatting process, Blackmagic disk speed tests and also the real world disk speed tests. That video uses a larger tests folder as well, for showing longer sustained read and writes. Here's some links to the stuff used in the video. I have included links to the Samsung 990 Pro, only because some people will still be interested in it regardless of what I say. Unfortunately, some KZbinrs are using this SSD and cherry picking the results and not being totally honest about its problems or that of the Mac's cache. Me personally, I would avoid the Samsung 990 Pro for use in external enclosures. Both the WD SN850X and Samsung 990 Pro are the ones WITHOUT the built in heatsinks. Use my promo code David15 for a 15% discount off anything that you buy on the Acasis website. AMAZON ASSOCIATE DISCLOSURE: I am an Amazon Associate. My Amazon links are Amazon affiliate links. I earn money from qualifying purchases when you use my Amazon affiliate links. OTHER EARNINGS AND COMMISSIONS: I also earn money from other product links within my video description. For the sake of clarity and for the avoidance of any confusion, assume that I earn money from commissions from any and all links that I have within my video description. You can also send me a coffee donation via PayPal if you found my video super helpful: www.paypal.me/DavidHarry My Amazon pages with videos and product links: Amazon USA www.amazon.com/shop/davidharry Amazon UK www.amazon.co.uk/shop/davidharry If you would like to help my channel please use my global Amazon Affiliate links. I will be paid a small commission for anything you buy from Amazon when using these links. These commissions don't cost you any extra but really do help me to buy gear for my productions and to review: Amazon USA: geni.us/Amazon-USA Amazon UK: geni.us/Amazon-UK Amazon Deutschland: geni.us/Amazon-Deutschland Amazon France: geni.us/Amazon-France Amazon España: geni.us/Amazon-Espana Amazon Italia: geni.us/Amazon-Italia Amazon Canada: geni.us/Amazon-Canada Contact for product reviews: [e-mail] www.DavidHarry.com I’m David Harry. Thank you very much for watching this video, take care and goodbye now. Cheers, Dave. #thunderbolt5#thunderbolt4#thunderbolt#thunderboltssd

[@philfyphil]

Excellent work David, thanks.

[@DavidHarry]

Thank you and you are very welcome 👍 HAPPY NEW YEAR. Cheers, Dave.

[@aaronlee6821]

David, thanks for the great video. You mentioned that your channel isn't about testing these SSDs, but I just subscribed to your channel because that's what I'm primarily interested in. This video plus your previous videos. It was based on one of your previous videos that I purchased the Acasis TB4 enclosure with fan, and the WD SN850x drive, and I must say I'm very pleased with the combination. Thanks for your hard work in comparing things, and for the heads-up on the Samsung drive, which I might have purchased if not for first seeing your video. And now for this video, it's a great resource for me. I was well pleased to see that indeed I can expect to double read/write speeds once I move to TB5, from my present TB4 hardware. And I do plan to purchase the same TB5 Acasis enclosure that you used in this video, when that time comes, unless something better comes out before then. I'm well pleased with the TB4 Acasis enclosure, especially with the fan. One comment, regarding your statement about the WD cache and perhaps being above a TB. Well, of course the cache itself cannot be that large as a TB of RAM would cost way more than the drive. So, I'm assuming you're referring to a combination of the reads/writes and pushing data into/out of the flash, and that if there's any filling up of the cache that's happening, it won't completely fill up until about a TB of data is transferred. I'm not sure, but perhaps it'll never fill up, if it didn't fill up in that test you did. Might be interesting to do some test about that. Anyways, I really appreciate your real world SSD tests. That's what's important to users. The best specs in the world mean nothing if they don't perform up to those levels in the real world. Keep up the great videos.

[@DavidHarry]

Hi, Aaron. I've just seen your comment, however, I will have to respond tomorrow/later today. I'm in the UK and it's now 3:30 in the morning and I should have got to bed about 5 hours ago 🤣 Cheers, Dave.

[@DavidHarry]

Alright, Aaron. On the point of the cache. The cache is the fast NAND and not the DRAM buffer. In this instance and it's been verified by someone else in the comments. The cache is indeed over 1TB, which is what I'd noticed in my tests. @KoenKooi has verified in another comment to this video that the 2TB has 600GB. As the cache does scale, this means that the 4TB has 1200GB, which is what I'd noticed in my tests when I said it was at least, or over 1TB. As for the DRAM, @KoenKooi also confirmed that this is 2GiB. The DRAM is what will be buffering the data and the cache is used for storing and moving the data to the rest of the NAND. So between them, this is why the WD is faster with the larger writes compared to the Samsung. The Samsung may be slightly faster in shorter bursts but that wont really be appreciable. The value and usefulness is in the larger transfers, especially if used as a storage drive. For anything else, such as an OS drive, their I/O speeds and latency will be very similar, although I suspect the WD is slightly better. Although, as an OS boot drive, I doubt there's anything between them in practical usage, they are both going to be great boot SSDs as long as the Samsung can be kept cool. Although, my personal preference is the WD as it's easier to cool and has more than twice the cache (that I've noticed in my tests). I'm glad that you've found my videos useful and that they've helped you to put together a drive that you are happy with. I know my typical test videos are long and boring. However, for someone who is serious about putting a drive together and not just speed clicking through KZbin videos looking for eye candy. I hope all the extra info in my videos is useful toward a purchasing decision. Plus, I personally can't stand watching KZbinrs using pointless graphs that don't show any evidence of the numbers. Or those who cherry pick results. I know most people want to see shorter videos and can't be bothered with all the detail. Well there's enough channels doing that for those types of viewers, even if the information isn't entirely accurate. I'm just happy that there are people out there, like you, who are interested in the "boring" information and the evidence that supports it. As for you subscribing because of the SSD videos, thanks 👍 My videos are usually general tech stuff to do with Macs, phones, tablets, gaming, video editing etc. I even do things like showing people how to apply a screen protector to a phone, or how to connect headphones to a tablet. So there will be a lot of my videos that you definitely won't be interested in. However, my titles are usually very descriptive of what the video is about, so you can just ignore the stuff that you know you wouldn't be bothered with. Anyway. All the best for the coming year and I hope it's a happy, peaceful and prosperous one for you 👍 Cheers, Dave.

[@aaronlee6821]

@@DavidHarry Thanks for your reply, David. I see from your reply that I had a basic misunderstanding of NVMe cache technology. I'm very familiar with CPU cache technology, as I've written lots and lots of code for ARM CPUs where my software needs to interface with the cache in order to optimize performance. So, it never dawned on me that the NVME would have fast NAND cache, rather I just assumed all the cache was DRAM. Quite ingenious of the designers to do that, and yes, it now makes perfect sense that the fast NAND cache is indeed 1TB. I'm happy that I could learn something new and very useful from you. Regarding your comment, "I know my typical test videos are long and boring.", not at all. I'd gladly spend 20 minutes or more to gain knowledge of hardware and make a smart buying choice than to not watch the video and make a wrong choice. Not only in wasting time/money buying the wrong product, but in using the product. Based on your real world speed tests, with the amount of copying huge files that I need to do, the time to watch one of your videos can easily be made up for in a single day of copying such files. And over the course of the lifetime of the product, I'll save countless times over the time I spend on your video. Thanks so much for doing what you're doing, even if some of your audience might find it boring. Even if your general content isn't useful to me, I will keep an eye out for your future videos regarding external SSDs, and anything Thunderbolt related. Happy New Year

[@joehernandez4534]

Amazing content and information. Just subscribed! Great work!

[@steelstunners1862]

I’ve been waiting for this vid. I bought both these enclosures yesterday

[@DavidHarry]

Awesome. Hopefully the video was useful for you 👍 Cheers, Dave.

[@pureheartvisuals]

Great stuff Harry… definitely would like to see the WD SN850X inside the TB 5 Enclosure.

[@DavidHarry]

Hi. You are very welcome. BTW the SN850X was inside the TB5 enclosure for the second set of tests. HAPPY NEW YEAR. Cheers, Dave.

[@rachidfinge8601]

Thanks David, this confirms my thinking - Thunderbolt 4 would be *just* too slow to get the maximum out of my 850X, so purchased the Acasis TB501 to get the maximum performance soon.

[@DavidHarry]

Hi. Yes, Thunderbolt 4 is great and you do get great speeds with the right SSD and enclosure. However, for absolute speed when speed is important, Thunderbolt 5 is definitely the way to go. HAPPY NEW YEAR. Cheers, Dave.

[@JayfkProductions876]

Thank You for making this video 👊

[@DavidHarry]

Awesome, and you are very welcome 👍 Cheers, Dave.

[@jacegreen3696]

The Acasis enclosures are great value for money. Compact and very well built with good quality cables. I recently bought 2 of the TBU405 Air enclosures that are only TB3/4 compatible, not USB for $60 each and a EC-7252 dual bay SATA SSD RAID enclosure for $25. No complaints. Your vids have been very informative and helpful. Thanks.

[@DavidHarry]

Hi. Yes, I totally agree. The Acasis enclosures are great and there's a lot of choice as well between passive and actively cooled ones. Some people may think that I'm a shill because Acasis send me these things for review. However, the only reason why they do is because they seen some videos of mine with Acasis enclosures that I had bought and they liked the videos and got in touch to see if I wanted to test more. So yes, I was a massive fan even before Acasis got in touch with me. I will see about getting their cheapest enclosure to test, that should be an interesting video. Thanks for your comment 👍 and I'm glad that you've found some of my videos interesting. HAPPY NEW YEAR. Cheers, Dave.

[@tombyrer1808]

Thanks for the test! BTW, I recommend to use the CLI to test transfers; sounds like the OSX UI is not 100% accurate timing wise.

[@DavidHarry]

Hi. You are very welcome 👍 Yes, using something in terminal would be very accurate. However, as I've said in another comment, the stop watch is easier to see when the video is viewed back on a small screen like a phone or tablet. HAPPY NEW YEAR. Cheers, Dave.

[@tbirdvet]

Good review. I just tested the Samsung 990Pro in my Trebleet TB5 enclosure and I got the same speeds numbers you did with the WD 850X. in the acacias. TheTtrebleet cost is less than the Acasis..

[@DavidHarry]

Hi. I'm glad you liked the review 👍 And awesome results with your TB5 drive. HAPPY NEW YEAR. Cheers, Dave.

[@rafograph]

Really interesting findings

[@DavidHarry]

Yes, I was really interested to find this out myself. HAPPY NEW YEAR. Cheers, Dave.

[@rafograph]

@DavidHarry thank you! Happy new year!

[@greatpix]

I'm just getting into using Thunderbolt enclosures (thanks for the videos) with my new Mac Mini and decided to go with two enclosures with 2TB each. One is fanless and has a Teamgroup MP-44L and the other has a fan and uses the WD Black SN7100. The WD will be for /Home folder and the Teamgroup for general storage and cache disk for some programs. If it gets too hot using as a cache I'll use the internal disc of the Mac. It may end up on a Windows PC.

[@DavidHarry]

Hi. I'm sure you will get results but just be careful with the temperatures. I wouldn't personally recommend moving your Home folder externally for a few reasons, the main one is that you now have two potential points of failure instead of one. However, if you do, wherever possible use the Mac's internal storage for any constant activities, such as file caching for certain programs if you can. This may help with heat issues on the external. HAPPY NEW YEAR. Cheers, Dave.

[@rorywalters1614]

I’m curious if the ‘cache’ issue was the internal SSD throttling. Macs are too conservative with their fan policy, especially with MacBooks, the fans don’t even spin when it’s not hot enough, at least the Mac minis’ would spin at a very low speed upon startup. I have a habit of turning the Macs‘ fans all the way up to 100% percent when doing heavy workloads (video editing, gaming etc.), I run my main OS off an USB4 SSD and the enclosure got a fan which runs pretty loud, so I don’t mind the Mac mini itself runs loud.

[@DavidHarry]

Hi, Rory. No, the cache issue on the Mac is actually the cache and not thermal throttling. I have done these tests a stupid amount of times and it's always the same. Even if I come from a cold boot and the Mac is really cool, if I run the test immediately after a cold boot the same issue happens. I know it's not scientific but the top, back and bottom of the Mac don't even get warm after these tests. Even my thermometer shows really cool temps. I did have a sensor monitor application to measure CPU, GPU, storage and RAM temps etc. but I found that to be unreliable, although, it wasn't showing any heat issues with the storage either. HAPPY NEW YEAR. Cheers, Dave.

[@rafapelikula7482]

Hi David! Here again to ask. I'm going in between Pc and Mac, is Exfat slower? for R/W speeds? Thanks

[@DavidHarry]

Hi. ExFAT will be slightly slower for writing on the Mac, not sure about Windows. Read speeds are very similar to APFS. However, it is the best format for sharing between the two. HAPPY NEW YEAR. Cheers, Dave.

[@ThirdEyePix]

Thanks again for such a real-world helpful video! I have a 2TB Mac Mini M4 Pro and have just ordered an Acasis TB501 plus WD 4TB SN850X after watching your 50-minute How To Build video. Regarding the potential Apple M4 Pro cache issue, is it something which Apple could solve in a future OS update? Or is it a physiacl hardware issue? < I ask this in ignorance!

[@DavidHarry]

Hi. I'm glad the videos have been useful and I'm sure you will absolutely love your new drive 👍 The cache issues I'm talking about with the Mac are hardware and not really something that is usually changed with firmware. However, your Mac Mini M4 Pro's storage is twice the size of my MBP's storage and the cache will be at least twice the size. So you may not even see this problem. HAPPY NEW YEAR. Cheers, Dave.

[@HoundDogMech]

You guys talked me into buying a Expensive Thunder Bolt Enclosure and 2 TB Samsung 990 EVO SSD. But when I connected my old 10MB/s Sabeant SSD to the New M4 Mac Mini Base model and told Davinci where those Video files where The Time Lines ran perfectlly well no Lagging even the audio was Perfect ounce it Cached.

[@DavidHarry]

I don't know who convinced you to buy a 990 EVO but it wasn't me. HAPPY NEW YEAR. Cheers, Dave.

[@nikolaiizotov6063]

I am glad that you paid attention to the internal architecture of the SSD and were able to correctly convey it to your audience. As a PC owner, I am more interested in the capabilities of the RAM disk and unfortunately I recently learned that those incredible speeds that it allows you to achieve (50 ... GB / s) are mercilessly limited by Windows Explorer at 5 GB / s, this limitation also applies to SSD drives with a PCI Express 5 v 10 interface - 13 GB / s. Do you know if such restrictions exist on the Mac platform? This indicator cannot be measured using standard programs since they work bypassing Windows Explorer, but what about the Mac platform? And even if they are not on the Mac, then to measure the limits of the Mac Explorer, you will need a specialized RAM controller that creates an ultra-fast SSD drive using RAM strips as a drive (PC platform). There are similar devices, but they are extremely rare and are used in servers, but most often servers use RAM directly and specialized software that works with it directly, bypassing the operating system.

[@DavidHarry]

Hi. I've never tried setting up a RAM disk on Mac and I'm not sure if there's any way to do this on Apple Silicon. There will have been for X86 Macs, although, I don't have enough RAM anyway. HAPPY NEW YEAR. Cheers, Dave.

[@nikolaiizotov6063]

Hello. I don't have a Mac computer, only an iPad and (a powerful PC). Trying to answer my question empirically is really a serious problem. I'm not sure that even Apple support is competent, since it's very specific. But you can try asking them a question. Happy New Year and good luck in your creative work.

[@darcsentor]

Great stuff, really interesting to see how these enclosures perform. Any chance of doing some random IO and latency tests? I’m looking at upgrading my Mac, and don’t know whether to destroy my wallet and get the largest internal SSD possible or attach some thunderbolt drives.

[@DavidHarry]

Hi. I'm glad you liked the video. I may do some I/O and latency tests at some point. However, as my videos are aimed at typical storage and not for OS/booting or database stuff, I just concentrate on the overall speed of the data dumps. However, if you are thinking of using an external for your OS, I would imagine that there will be a slight hit on I/O and latency with the overhead of the PCIe/Thunderbolt tunnelling/bridge in the enclosures chipset but this in all fairness is likely to be unnoticeable. If you are interested in using something based on I/O and latency for OS/booting or databasing etc. I would totally recommend the SN850X. I'm still unsure of its DRAM size and type but the SSD is super quick for latency/I/O sensitive activities. Plus, if you are using macOS, the DRAM is a big deal as neither the OS or SSD use HMB (Which I think is a MS/Windows thing anyway). I've used the 4TB SN850X as a boot drive for my Apple Silicon machines and even with the "presumably" faster storage on the M4 Max. There was no practical difference using the SN850X for booting, it was super fast and responsive. You will miss out on a few things with external macOS booting but I don't personal use things like Apple AI, Apple Pay etc. on my work machines, so booting external isn't an issue. HAPPY NEW YEAR. Cheers, Dave.

[@darcsentor]

@@DavidHarry Thanks for the reply. I am software dev. I have a thunderbolt 4 ssd for onedrive etc. Got interested yesterday after your video and tested code compile on it. Much to my surprise it was a similar speed to the internal SSD on the Mac mini. biggest problems I've had to date have been getting everything attached to the M4 Mac mini stable. Had a Time Machine back up connected via USB that managed to corrupt at one point. Little hesitant on external drives now. I'm used to either throwing an extra M2 drive onto the motherboard or a PCI card. Ended up 3-D printing a case for my Mac mini, to hold it and all the Thunder bolt docks and external SSD together in my rack.

[@asafblasbergvideographer]

Are you copying data from the external SSD to the internal SSD on the Mac? That was the only thing I Was confused about thanks,

[@DavidHarry]

Hi. Yes, both sets of tests are going to and from the Mac's internal drive. Cheers, Dave.

[@asafblasbergvideographer]

@DavidHarry OK thanks for the clarification. The TB5 speeds are impressive; my only concern is if you plug in a Thunderbolt dock will the speeds decrease due to bandwidth limitations? On my Mac Mini I have a Thunderbolt dock that attaches two 4K displays. Thank you :)

[@DavidHarry]

@@asafblasbergvideographer Ahh, funny you should ask. I have recently been sent a Thunderbolt 3 dock and Thunderbolt 4 dock. Unfortunately, no Thunderbolt 5 dock so far. Aside from demonstrating the docks and showing their ports etc. One of the videos I will be doing is to show how the Thunderbolt bandwidth, no matter what version of Thunderbolt, will reduce when using video outputs. With Thunderbolt, video always has preference over the data bandwidth. So for instance, if you attach two 4K monitors in 60FPS, the bandwidth for your data (SSDs) will be massively reduced. I will also do a video explaining why you never get 40Gb/s for your data bandwidth, even if the only thing attached is a super fast SSD. The problem I have is that I am very late on about 10 videos, so need to clear them before I can do more geeky stuff. BTW. If you have your Thunderbolt dock on one port of your Mac and the Thunderbolt SSD connected to a different port, you will get full speed on the SSD no matter what the dock is doing. The Macs do use independent Thunderbolt busses, so no data/bandwidth sharing/splitting across the different ports of the Mac, they are all independently 40Gb/s or 80/120/Gb/s. Cheers, Dave.

[@asafblasbergvideographer]

Excellent, thank you so much

[@estevegraells]

Acasis about 30% more expensive than ugreen with active cooling included. Anyone knows why?

[@Jasmine-Law]

Check this out - it's all about the chips! ACASIS goes with Intel's Thunderbolt chips, which yeah, cost more, but they're way more reliable. Ugreen? They're using TSMC's ASM2464 chips to save some cash, but man, those things run super hot and keep dropping connection. Not worth the headache if you ask me.

[@DavidHarry]

Hi. Yes, the chipset does make a difference. Also, with TB5 as far as I know there are only Intel chips so far. HAPPY NEW YEAR. Cheers, Dave.

[@DavidHarry]

Hi. I don't know why but I can say that the Acasis enclosures are great, although, I've not used Ugreen. HAPPY NEW YEAR. Cheers, Dave.

[@filipr6070]

I have an Acasis TBU401 + Samsung 980 pro and it runs hot. It is not recommended as bootdrive.

[@DavidHarry]

Hi. Yes, I have a few Samsungs that run uncomfortably hot and you are absolutely correct, they are definitely not recommended as boot drives when inside an enclosure. The only time I would use Samsung's that have DRAM, is in a Windows build where you can use much better cooling. I've used some of the cheaper Samsung's that don't have DRAM and they are usable, although, not as fast for booting or for sustained writing as a storage drive. I'll do a video soon with a mid-priced Samsung SSD 980, which is OK for temperatures in an enclosure as a storage drive but not the best as an OS/boot drive externally. HAPPY NEW YEAR. Cheers, Dave.

[@TheDanEdwards]

If you want to avoid the latency of the Finder, you can always use the _cp_ command in a Terminal. Use the _time_ command in conjunction with a Terminal command to get a timing of how long it takes.

[@DavidHarry]

Hi, Dan. Yes, that's a great option and suggestion. The only problem with using terminal is I have to zoom in too much in post and the picture doesn't look so good. Using the stop watch is just so that people can easily see the time. Plus, most people view back on small screens like phones and tablets, so I have to be careful how small stuff is on screen or how blurry it goes when blown into in post. HAPPY NEW YEAR. Cheers, Dave.

[@TheDanEdwards]

​@@DavidHarry Understand the problem of font sizes for video. Note that one can make the font quite large in Terminal. Give it a try!

[@greatpix]

I hope you don't mind my commenting on this but you're coming out sort of greenish grey in this video. IMO all of your videos could use with some warming up. I know YT maps colors its own way so maybe that's it. If this were the old film days, before I retired over 20 years ago, I'd slap a 1/4 CTO over the hot lights and call it a day. You've got it easier nowadays.

[@DavidHarry]

Hi. I've been doing audio and video post for over 35 years and while I'm not a DP, the colour and lighting is fine and everything is balance 5200K. However, my rostrum camera can sometimes have issues with contrast (blowing out a bit or being slightly crushed in the shadows) due to the nature of the subject matter and average frame exposure. HAPPY NEW YEAR. Cheers, Dave.

[@greatpix]

@@DavidHarry This is my 3rd attempt at a reply, I keep hitting the Esc key by accident and losing everything. My typing skills aren't what they were even in my late 60's. I'm in my mid 70's now and spent the last two years in and out of the hospital with severe health issues. I kid people that I was born holding a MacBeth Color Checker. I've been color picky since the 60's. I managed pro camera stores for over 27 year, been a tech writer for several online magazines, been a pro glamour photographer and even shot for Playboy's website, and other things like a founding member of the first Apple Computer club. I studied film in university and used to go up to USC's Film School for special events and met some interesting people, like a young George Lucas. This is my thinking... People buy warm light light bulbs over bright white light bulbs because they give a more pleasing skin tone. I think your videos would benefit by going a little warmer on your lighting, 4800K might be enough, maybe warmer. If you were doing videos where you had to reproduce the color of a piece of artwork then accurate color comes to the fore. I hope you'll feel like experimenting a bit with this and see what develops.

[@DavidHarry]

@@greatpix Hi. Don't get me started about hitting the wrong buttons 😂 I couldn't tell you how many times I've accidentally deleted long comments or forgot to post them 🤣 Sorry to hear about your health issues. My background is in film, TV and music. When I say film, I mean shooting digitally with only a bit of 16mm. I have been head of post for two UK features and contributed in a number of ways (post) with TV and DVD content. I personally don't like a warm look and my main Sony shoots slightly cold (which I prefer) when balanced with the lights. My approach to KZbin is a traditional TV workflow, shoot, edit, print (video). I do not do anything in post other than an edit and audio mix and just use what the camera has recorded. If I have a noticeable issue, details in shadows going missing for closeups from my rostrum shots, I will raise the shadows in post. As a one person operation you have to be pragmatic with your time for a KZbin video, especially when producing a lot of them. This is why I don't have an issue with the picture tone. My videos are clear as far as picture and sounds are concerned and that is the main requirement for these instructional type videos. If I were involved in a TV or cinema production, there would be someone specific doing a grade, it's not something that I would get involved in unless it was for a "look". I most certainly wouldn't get involved with a technical grade or colour/scene balance as I'm not good enough for that and someone who does that as their only job is way better suited to it. If I had a number of people noticing issues, I would take a look at it. However, after producing over 1500 video across multiple KZbin channels, you're the only person who's ever commented on the picture. So like I said, pragmatism rules the day for me and I cut my cloth accordingly. Have a great New Year 👍 Take care. Dave.

[@tomsun3159]

Hi David, its clearly the same problem why Fast NVME SSDs have an DRAM Cache the speed of the NAND-Chips is limited this is also valid for the NAND-Chips used in the apple products, they will use a certain amount of unified memory as a cache if this is eaten up it will throttle (same issue in the comparison of Samsung vs. WD Cache-Size matters) So it is a problem by design with the decision not to use standard NVMEs. Its the same issue with cars having an 1.0 ecoboost or an 3.0 R6, but on the long run you won't have fun with the downsized engine with its design faults like a timingbelt running in engine oil, yes it will work but not for long. The real question is depends the amount of unified storage used for caching on the the amount of unified memory, is there a difference between 16 GB and 128 GB of unified memory (and the sizes between)? Regarding the testfiles, of course you take testfiles from your personal workflow, i assume the average filesize is quite big (videofiles?) Does the results vary if using lots of small files (e.g. source code of big projects, i would think of multiple sets of linux sourcecode to reach an comparable size, should be thousands of files of small size) After all TB5 has clearly double the speed in the copying activities, moving a large bunch of files, but does it have a real benefit when working on them (so if you do not the whole project on the internal drive e.g. rendering a large video not on the internal drive)?

[@DavidHarry]

Hi. Yes, Apple should have just used standard SSDs and gone with something that had high DRAM. I'm still unsure if the "DRAM" is separate on the SoC or if it's using the unified memory. I only say this because if it were using the unified memory, I would have thought that it would be dynamic and scale with activity but it doesn't seem to do that. However, the DRAM doesn't even come into it once the NAND cache has been depleted. Either way, I think Apple have short changed people with the SSDs/storage, especially considering the absurd prices they charge. I can't comment on your car engine analogy. Aside from understanding the basics of an internal combustion engine, I don't know anything about them. Plus, I don't drive 🤣 As for file size. Yes, these are my own personal files. They are video videos, music, pictures, project files and some old web site backups. Most people will using similar files on storage and media drives, so my overall speeds are aimed at the most type of use these drives will get. My small web files only number a couple of thousand in the tests with big folders. However, I also use the SN850X and TB4/TB5 for booting macOS and it's indistinguishable from booting and running internally. There are some Apple limitations doing this, AI, Apple Pay etc. but this doesn't bother me as these are work machines. I don't just have my "Home" folder external, it's the whole OS. I would not recommend moving the Home folder as it's problematic for a number of reasons and splitting your OS activity across two drives is just dumb in my opinion and you are asking for trouble, which you will inevitably get. You also introduce two potential points of failure instead of one when moving your Home folder externally. As for the speed and if you need it. This is down to individual use. Yes, you need it if booting externally but as I've proven with my recent SATA SSD video editing test, most people don't even need to go above the speed of an external SATA SSD, about 400MB/s for most types of video editing. For me personally, I do loads of high bandwidth stuff, even using ProRes 4444 XQ but mostly 422 HQ and I will also go up to 8K 120FPS but mostly 4K 60FPS. However, hardly anyone does this and my use is far from typical. Anyway, like I said, the use of super fast external SSDs is down to the individual to justify. Also. Going from TB4 to TB5 doesn't necessarily scale. This all depends on the SSD used and if it can take advantage of the faster TB5 data bus. I will do a video at some point to show that some SSDs wont be much faster with a TB5 enclosure compared to a TB4 enclosure. HAPPY NEW YEAR. Cheers, Dave.

[@cassio2999]

fooooorrrrty

[@DavidHarry]

🤣👍

[@KoenKooi]

For the WD cache size, the Techpowerup SSD database reports 2GB of RAM and 600GB of pSLC cache for the 2TB SN850x, so if that scales linearly, the 4TB version would have 1.2TB of pSLC cache, but still only 2GB of RAM. If I run smartctl on my 4TB SN850x, it reports a 10% reserve (400GB), but I don’t know how honest the drive is when reporting such things :) And a happy end of 2024!

[@DavidHarry]

Hi. Thanks for the info, this is something that I've been trying to get to the bottom of. I suspected that the fast NAND (SLC?) was around a terabyte for the 4TB. This was just something that I'd guessed after constantly moving multiple 500GB folders, as it dropped when writing the third pass of 500GB. So thanks for confirming that 👍 While I still don't know the cache on the 4TB 990 Pro, it definitely falls short toward the end of a 500GB dump, so maybe about 450GB. In any event, the WD is over twice the size, which is why it's much better for sustained data transfers and faster with longer transfers. BTW, I know you know that, I'm just mentioning it in case anyone else reads the comment. Now, as far as the DRAM, I'm still uncertain about that. I had also read that it was 2GiB and didn't scale with the capacity of the drive. However, I read elsewhere that it was more and it did scale with the capacity of the drive. None the less, 2GiB is easily enough anyway and the drive performs really well. HAPPY NEW YEAR. Cheers, Dave.

[@KoenKooi]

@ Happy New Year! From what I understand is that the SSD treats a portion of the TLC flash as SLC to get faster write speeds. During idle times it will ‘fold’ the cells back to TLC. The numbers I can find on the internet say that for the sn850x the pSLC write speed is 6gbyte/s, the TLC write speed is 3gbyte/s and the folding speed is 1.5gbyte/s. If the 4TB version provides 1.2TB of SLC cache, that means it’s actually using 3.6TB (only 1 bit stored per three level cell) worth of TLC cells, so if there is more than 400GB used, the cache will be less. I don’t know for certain if that’s how the sn850x works, but your observations fit :) But for the SSDs we qualified at my previous job the behaviour was completely different, so it depends on the controller and firmware used. Anyway, for my use case, photo and video, the SN850x works very well, no slowdowns when offloading a day worth of sitting in a hide, 30fps at 45MP and 8k60 video will quickly fill the cache :)

[@JayfkProductions876]

Avoid Samsung 990 like the plague 😂 their firmware issue was a massive L

[@DavidHarry]

Hi. I have seen some horror stories from Windows users using them inside their builds. I'm sorry I bought mine 🤣 Cheers, Dave.

[@JayfkProductions876]

@@DavidHarry it was one of the weirdest f ups on Samsung's part, where that issue would pretty much brick the drives, they said they fixed the issue but maaaan the risk scares me so I stuck with my 970's

[@aaronlee6821]

NAND flash is very complicated, due to the fact that it develops errors over time. The algorithms that deal with that are quite complicated, and need to be exhaustively tested. I would never ever trust a product that uses NAND flash and is dependent on algorithms/drivers written by a small company who likely doesn't have the resources to do it properly. I have lots and lots of horror stories from years past of products suddenly dying due to this very issue, where the software engineers just assumed that the NAND flash would last forever and had zero functionality for error detection/correction, or had very poorly implemented it. I assume with the huge size of Samsung that they have plenty of engineers capable of writing the drivers necessary for NAND, but the question is do they actually do it properly even though they have the ability. Poor management of the development process, including testing, can easily result in catastrophic results in NAND flash products. That said, I've used numerous Samsung NVMe drives, including a newer 990 pro in a PC, and have yet to find any problems. For smaller companies, I assume they're using NAND drivers/algorithms developed by a large company that has expertise in that area, but I'm always leery about those products due to my past bad experiences with NAND products. For NAND products, I'll only buy from top-tier companies, and only after checking reviews.