ADATA XPG Gammix S70 Blade 2 TB

[W1zzard]

At $300 for the reviewed 2 TB version, the ADATA XPG Gammix S70 Blade is one of the most affordable PCIe 4.0 SSDs available today. Despite its low price, it offers compatibility with Sony's PlayStation 5, and performance is comparable to the Samsung 980 Pro, WD Black SN850, and Corsair MP600 Pro.

Show full review

 

[Selaya]

Why is this great value? Is PCIe 4.0 really worth $50-100?
I mean I get that the PS5 is demanding this but still ...
(I wonder how much of that is actual necessity and how much is just marketing BS. I'm honestly inclined to believe that it is mostly the latter. And I wonder whether there's a way to test this, possibly by MacGyvering together an m.2 drive that can log data access/transfer rates or something?)

 

[Tardian]

The 2 TB ADATA XPG Gammix S70 Blade is 496 AUD and the ADATA 2TB XPG SX8200 Pro PCIe M.2 NVMe SSD is 399 AUD currently.

The Good:

• Very good real-life performance
• Competitive pricing (for a PCIe 4.0 drive)
• Excellent synthetic performance
• Large SLC cache
• Good sustained write performance for a TLC drive
• PCI-Express 4.0 support
• Compatible with Sony PlayStation 5
• Heatsink included
• DRAM cache
• Five-year warranty
• Compact form factor

The ugly

• Heatsink has very little effect
• More expensive than PCIe 3.0 drives with similar performance
• Thermal throttling even with heatsink installed
• Largest capacity available is 2 TB

Tardian notes that the reviewed product took 43.9 seconds in Adobe PS CC whilst the older product took 46.6 seconds.

Conclusion a 6% improvement for a 24% price hike is bad value IMHO. Look elsewhere is my recommendation.

 

[mechtech]

Nice review W1zz,

Some constructive feedback.......would it be possible to put up (SATA) & (NVMe) beside the drives? There is so many it's hard to keep track

Cheers

 

[iuliug]

Adata branded NAND means they can change it whenever they want to QLC like they did b4.

 

[Tomorrow]

Why is this great value? Is PCIe 4.0 really worth $50-100?
I mean I get that the PS5 is demanding this but still ...
(I wonder how much of that is actual necessity and how much is just marketing BS. I'm honestly inclined to believe that it is mostly the latter. And I wonder whether there's a way to test this, possibly by MacGyvering together an m.2 drive that can log data access/transfer rates or something?)

Great value compared to other PCIe 4.0 drives. In my country the S70 costs 317€. That's even cheaper than PM9A1 (980 Pro OEM version) and about 60-70€ cheaper than the cheapest non-OEM PCIe 4.0 drives.

 

[jcwbnimble]

I will never trust XPG products again. I fell victim to the switcharoo that they did with the SX8200 Pro drives. They submitted drives with high end components for tech reviews, suckered a bunch of people (including me) into buying the drive, then sold sub-standard drives with inferior components. Never again. Buy samsung, Crucial, Intel, ANYONE BUT XPG

 

[TheLostSwede]

I will never trust XPG products again. I fell victim to the switcharoo that they did with the SX8200 Pro drives. They submitted drives with high end components for tech reviews, suckered a bunch of people (including me) into buying the drive, then sold sub-standard drives with inferior components. Never again. Buy samsung, Crucial, Intel, ANYONE BUT XPG

News flash, everyone's doing it these days.

 

[kuro]

News flash, everyone's doing it these days.

Fair, but at least the tests on Samsung's show that there's at least a benefit of the swaps. ADATA's swaps are a performance decrease all around.

 

[Jism]

News flash, everyone's doing it these days.

You want premium you pay premium.

It's not like these drives coud'nt cope up with any daily type of workload really. It's just not suitable for heavy sustained loads, i.e video editting on the long run without performance tampering in.

Its good for a OS boot drive, loading up games etc. Thats about it.

 

[hojnikb]

Can we place bets, how long it's going to take adata to switch out the nand and/or controller for a cheaper version.

 

[Tsukiyomi91]

$150 for the 1TB model while still giving decent performance sounds good to me. still cheaper than Corsair's MP600 and Samsung's 980 PRO.

 

[GabrielLP14]

I will never trust XPG products again. I fell victim to the switcharoo that they did with the SX8200 Pro drives. They submitted drives with high end components for tech reviews, suckered a bunch of people (including me) into buying the drive, then sold sub-standard drives with inferior components. Never again. Buy samsung, Crucial, Intel, ANYONE BUT XPG

Well, what if you find out that a lot of companies do this? Samsung did this on the 970 EVO Plus (from Phoenix to Elpis as controller and from 92L to 128L as NAND), WD on the SN550 (denser flash), Crucial on the P2, MX500 and BX500, PNY on the CS3030, Patriot on the VPN100, etc.

 

[Dragokar]

News flash, everyone's doing it these days.

And everyone should be avoided at all costs if they don't change the SKU.

 

[jcwbnimble]

Well, what if you find out that a lot of companies do this? Samsung did this on the 970 EVO Plus (from Phoenix to Elpis as controller and from 92L to 128L as NAND), WD on the SN550 (denser flash), Crucial on the P2, MX500 and BX500, PNY on the CS3030, Patriot on the VPN100, etc.

I can understand changing components as newer tech comes along (more layers for NAND) or even if supply of one component is limited (cuz fake pandemic) as long as performance specs are maintained. I can attest to the XPG swap as NOT being of equivalent specs as what was tested by mass media. So no, not surprised by any of these tech companies playing bait and switch, but there are some who are more reputable than others. I fell for the hype, and couldn't afford the Samsung NVME drives I wanted (plus they were hard to get) so I bear the responsibility for my own purchase, BUT I don't forget or forgive so no more XPG or Adata products for me....ever.

 

[GabrielLP14]

as long as performance specs are maintained.

Yeah, in fact the Samsung 970 EVO Plus degraded as performance from 1500 MB/s to around 850 MB/s (obviously post SLC cache), even if the SLC cache has increased in size.