Wednesday, March 6th 2024
Sparkle Intros PCIe Gen 4 Quad M.2 Riser Card
Sparkle introduced the PCIe Gen 4 Quad M.2 Riser Card. This single-slot, full-height card is meant to be installed on a PCI-Express 4.0 x16 slot, which it wires out as four M.2-22110 slots with Gen 4 x4 wiring, and mounting bolt set to M.2-2280 by default. With the drives in place, the card is meant to be covered up by a metal shroud; there's no contact between this shroud and the drives underneath. Cooling comes in the form of two lateral-blower fans that guide airflow under the shroud.
The first (intake) fan is 60 mm in size, while the second one (exhaust, near the tail-end), is 50 mm. Both are controlled by a localized fan-speed control, with three speed settings that can be selected by a 3-way switch on the rear I/O shield. Next to it, are four LEDs that denote power/activity of the individual drives. The card is 28 cm long, and 12 cm tall. What sets Sparkle's card apart from other brands is that the card relies entirely on the PCIe slot for powering itself—unlike some other cards that use a 6-pin connector. The company didn't reveal pricing.
The first (intake) fan is 60 mm in size, while the second one (exhaust, near the tail-end), is 50 mm. Both are controlled by a localized fan-speed control, with three speed settings that can be selected by a 3-way switch on the rear I/O shield. Next to it, are four LEDs that denote power/activity of the individual drives. The card is 28 cm long, and 12 cm tall. What sets Sparkle's card apart from other brands is that the card relies entirely on the PCIe slot for powering itself—unlike some other cards that use a 6-pin connector. The company didn't reveal pricing.
13 Comments on Sparkle Intros PCIe Gen 4 Quad M.2 Riser Card
Yes I have tried 4
They've been making Intel Arc GPUs for a while now.
And ASUS has already released a PCIe 5.0 version, so no idea why anyone would release a PCIe 4.0 card now.
Sparkle also entirely failed to fix the most annoying thing about this design, which is that it's taller than full-height (see how the PCB sticks up above the slot bracket)? That makes getting the slot screw in there an annoyingly difficult endeavour. If the drive mounts were tilted at, say, 35 or 40 degrees from the horizontal instead of the current 45, the PCB could be cut down to full-height - although a bit longer.
The extra fan is useless and dumb, they've very obviously just shoehorned it in there because there's space on the PCB and they want to appeal to people who don't understand how anything works. The fact that the shroud is cosmetic (versus the ASUS design, where it's a one-piece aluminium heatsink) makes it pretty much useless.
www.gigabyte.com/SSD/AORUS-Gen4-AIC-Adaptor#kf
I think many who considers such products are planning to use 2 or 4 (overpriced) consumer SSDs in RAID0 to boost performance, even though makes for only short bursts of high speed and terrible reliability. While it makes for great click-bait YouTube videos, it makes little sense compared to using a single proper high-end enterprise SSD (like e.g. Kioxia CM7-R/CM7-V), which can run circles around such a setup. And this is just a standard PCIe card right? So it should work on any kind of motherboard, provided it has bifurcation?
I got to ask, because even though anything PCIe should in theory work on anything PCIe, I know there are exceptions when it comes to "exotic" hardware, as not everything behaves 100% according to spec. I don't have the maximum power draw of M.2 off the top of my head, but I've seen them as high as 11 W at least, so we're talking in the neighborhood of ~50 W for four of them. In a server/workstation build with airflow through it, a tiny heatsink with small fins will do that just fine, but in a typical desktop where this card would be placed below the GPU and probably have mostly static air, you would have to have a massive heat sink with long fins to rely on that passively. Or do what most motherboard and SSDs do now anyway, just slap a big metal blob on top of there, enough to let it absorb heat long enough to look good in benchmarks, but put any medium sustained load on it and it wouldn't be able to dissipate heat quickly enough and just throttle like crazy. :(
As you can see in the example Chaitanya provided in the post before you; if there is a tiny bit of airflow, a small heatsink can do just fine.