My stupid brain upon seeing it. [youtube=arnWU1sWqKw]

I suppose the fan would last a decently long time if it only spins when the SSD is really stressed, which probably wouldn't be that often.

Let's see improvements in random 4k reads

4th gen is more than enough for my needs for many years, hopefully by the time 4th isnt good enough the newer gens wont need a big cooler on top Id rather see bigger than 4tb and 8tb at affordable prices than just more faster ones

This means iCUE is needed for this SSD??? nooooooooope i'll wait for SABRENT/WD

YUP Combining several posts 1. 4GB gen 4 are coming down in price to close to what I paid for my first 1GB SSD. I have been looking forward to gen 5 but have been disappointed with what I read 2. does anyone really spend that much time on sequential read and writes? The speed noticed by the average user is in the random reads and writes. I am not seeing anything like generational improvement in that regard. Hence the GEN 4 is the much better value. 3. Heat on the NVME slot positioned directly beside the graphics card is already a problem ON MANY MB. I understand that graphics chips can run safely at 80 to 90 C and that the new default is to run up to that thermal to get the highest performance OC. However SSD tend to throttle at those temps. AM I THE ONLY PERSON ON EARTH THAT SEES THE STUPIDITY OR THIS ARRANGEMENT???????

10 GB/s @ 10K RPM, whoop!

mdk77777:

3. Heat on the NVME slot positioned directly beside the graphics card is already a problem ON MANY MB. I understand that graphics chips can run safely at 80 to 90 C and that the new default is to run up to that thermal to get the highest performance OC. However SSD tend to throttle at those temps. AM I THE ONLY PERSON ON EARTH THAT SEES THE STUPIDITY OR THIS ARRANGEMENT???????

Multiple boards have multiple NVME slots now and 5.0. You can get PCI NVME cards to move the drives away from the GPU, but you may be locked to using AMD if 5.0 is needed on multiple PCI slots.

Yeah, but they put the NVME 5 slot as close to the CPU as possible .....I understand trace length. Take the GIGABYTE X670E AORUS MASTER AM5 board As you say, 2x NVME 5 SLOTS and 2x NVME 4 Slots. Great quality board, what could be better? Well, the first NVME 5 slot that you can use is above the graphics card backplate....even with a cooler it going to see radiant heat from the backplate going upward from 60c to 80c on high end modern cards. The second slot will be UNDER the graphics card, so no high rise cooler can be used at ALL. The slots you might be able to use at the bottom (assuming your high end graphics card is not wider than 3 slots(many are) is not NVME 5 Just saying this tall cooler with NVME 5 is not going to work well with the vast majority of MB in existence.

mdk77777:

Yeah, but they put the NVME 5 slot as close to the CPU as possible .....I understand trace length. Take the GIGABYTE X670E AORUS MASTER AM5 board As you say, 2x NVME 5 SLOTS and 2x NVME 4 Slots. Great quality board, what could be better? Well, the first NVME 5 slot that you can use is above the graphics card backplate....even with a cooler it going to see radiant heat from the backplate going upward from 60c to 80c on high end modern cards. The second slot will be UNDER the graphics card, so no high rise cooler can be used at ALL. The slots you might be able to use at the bottom (assuming your high end graphics card is not wider than 3 slots(many are) is not NVME 5 Just saying this tall cooler with NVME 5 is not going to work well with the vast majority of MB in existence.

The Motherboard NVME cooler between the CPU and GPU is huge on that board, maybe twice the size of the stock corsair MP700, any kind of airflow from a CPU cooler or case flow will be much better then that microscopic corsair fan. GPU backplates are very bad at transferring heat to surrounding air, many do not even have thermal pads to the PCB and many others are to thin to spread the heat out into the entire backplate and none of them have any kind of fins on them. My Gigabyte NVME with a tiny 2mm high cooler is sandwiched between a stock AMD direct 6900xt and a 5800x with a Noctua aircooler, It does not get much worse then that, I am running out of disk space before temperature is a problem. The 3 slot NVME cooler plate under the GPU is also massive, hot air from the GPU will hit it but I am sure the huge size it good enough. You could add a little thermal paste between NVME cooler and chipset cooler, to transfer a little extra heat onto the chipset heatsink, but i doubt it is needed.

some points: 1. both the new AMD and Nvidia generation of graphics cards allow the GPU to run up to an insane temp before throttling back. This is by design and the new standard. I have a 7900x and it got really toasty really fast, and yes it made my NVME get warmer than I like. I solved the problem by adjusting the fan curve on the card to keep thermals in line....not for the card, but for the NVME. 2. an air cooler actually assures that there is cooling / air movement around this area. People get in trouble more when you use an AIO, and base their case fans off of the CPU temp. This can lead to not enough air flow around the memory, VRM, and the NVME. This is the worst case for temps, not your air-cooler. Finally, I agree with you regarding the covers for the current gen 4 NVME. However, the concentrated heat of the NVME 5 is the problem at hand. Concentrated heat density requires more than passive cooling ....no matter how massive....it is a heat transfer rate problem, not total heat. Hence all the new GEN 5 you see require heat pipes and or a cooling fan. The point of my post was that MB makers will need to start accommodating this if NVME 5 SSD are to take off. The standard passive cooling covers WILL NOT WORK. They would have launched months ago if it did. PS, there is a fan less technology that is being rolled out for notebooks in the coming year. It looks like it has merit, perhaps that will be the solution. https://hothardware.com/news/frores-airjet-cooling-tech-aims-to-get-rid-of-laptop-fans But I doubt it will be cheap...spending $100-200 to help improve the cooling of a $3000 laptop is one thing. It will be awhile before it is cheap enough to cool a $300 SSD. Hence the problem the industry faces...yeah we can hit the new GEN 5 speeds, but not with passive cooling that is cheap and flat.

kx11:

This means iCUE is needed for this SSD??? nooooooooope i'll wait for SABRENT/WD

I got rid of Icue in favour of SignalRGB and if i need to control my fans its just a few quid a month subscription to save the headaches of Icue/Msi center

mdk77777:

Concentrated heat density requires more than passive cooling ....no matter how massive....it is a heat transfer rate problem, not total heat. Hence all the new GEN 5 you see require heat pipes and or a cooling fan. The point of my post was that MB makers will need to start accommodating this if NVME 5 SSD are to take off. The standard passive cooling covers WILL NOT WORK. They would have launched months ago if it did. PS, there is a fan less technology that is being rolled out for notebooks in the coming year. It looks like it has merit, perhaps that will be the solution. https://hothardware.com/news/frores-airjet-cooling-tech-aims-to-get-rid-of-laptop-fans

The heat concentration is not high at all, the max total power of the NVME slot is probably less then 10 watts and a average SSD controller size is around 1.8cm X 1.8cm. (Samsung 990) I have an old AMD athlon 1700+ here on my desk, TDP 50W in a size of 1.1 X 0.8cm, can be passively cooled, not easily though, but still half the size and 5 times the watts. A Ryzen 5600G is 180mm2 that is around 1.35cm X 1.35cm, that is easy to passively cool with a dumb and cheap aluminum cooler at 35w or even 45w. The airjet gismo thing does not look impressive for anything higher then 5 watts. It uses 1w to cool 4.25 watts of heat, so my new company laptop with a 15w CPU would need 5 of these modules using 5 extra watts of power to move air, a fan would use 1.9 - 2.5watts, so less then half the power and probably a lot cheaper.

Yeah, you are still confusing total heat with the required RATE of heat transfer. A cast iron radiator has a huge heat capacity. However you cannot cool a chip with iron because the RATE of heat transfer is not high enough. You could literally TIM an iron radiator to a modern CPU, and you would see heat throttling. The TIME before the SSD hits a critical temperature is the problem, not the total heat energy that needs to be dissipated. Heat pipes help this problem...Fans help this problem by breaking boundary layers, increasing the speed of heat dissipation. RE: your old chip cooling solutions Old technology is actually much more robust due to this. Car manufactures do not WANT to go to new nodes because of the price, but also because the environmental factors are greater (the newer chips cannot be allowed to get as hot) The same with fighter jets....yes if not broke don't fix...but also the newer chips are less robust ( not able to take the extremes and changes in temperature) I guess we will have to disagree for now. I don't think you will see 35 watt requirements for NVME 5 However you will see SOME other solution than the standard passive aluminum covers currently used. https://www.tomshardware.com/features/phison-e26-ssd-preview-pcie-5-ssd/2

Man the first company that focuses on 4k randoms if they do manage a significant gain they will monopolize among the enthusiasts , as far the price is not too high over the others .

Kaleid:

Let's see improvements in random 4k reads

Yep. I'd rather have maxed out gen 4 speed and 200 MB/S 4KQ1T1 than maxed out gen 5 speed and a small bump to 4KQ1T1. If you follow SATA 600 to PCIe 4X4, the percentage gains for sequential are FAR outpacing 4KQ1T1. Optane was cool, but just about everything outside of the endurance and 4KQ1T1 was terrible.

Venix:

Man the first company that focuses on 4k randoms if they do manage a significant gain they will monopolize among the enthusiasts , as far the price is not too high over the others .

Well, they need to nail the stuff Optane failed on. Optane failed on price to capacity ratio, total capacity and no compelling products in the M.2 2280 form factor. 4KQ1T1 alone is not enough.

So we gonna need active cooling with gen5? Just lovely but you gotta increase that load times so its worth it. 😀

mdk77777:

Yeah, you are still confusing total heat with the required RATE of heat transfer.

No. Total heat and size of hot surface is heat concentration, just like you wrote about earlier.

mdk77777:

However, the concentrated heat of the NVME 5 is the problem at hand. Concentrated heat density requires more than passive cooling ....no matter how massive....it is a heat transfer rate problem, not total heat. Hence all the new GEN 5 you see require heat pipes and or a cooling fan.

OK, perhaps concentration was a poor word choice my part. Reading up on the controller, the REP(President I think) from Phison states that it can tolerate 125C. However the memory chips are optimal at running bellow 50C and shut down at 90C. So, the heat intensity generated by the controller if spread over the same heatsink as the memory, you have a mis-match. The very high TEMPS (not total heat energy...temp does not equal energy) generated by the controller must be dissipated quickly. Rapid dissipation of TEMP from a source(even if low total heat energy density) = Heat transfer rate problem Again, all the models of the Gen 5 I have seen pre-launch have a fan (one way to increase the RATE of heat transfer) Perhaps there is another solution i.e. perhaps small heat pipes only attached on controller and not memory But you still end up with something more complex and more expensive than the fan solution.