1TB Class Performance Testing
We pulled in a number of popular 1TB NVMe class products to compare to the new Corsair MP510. Four models really standout starting with the low-cost Intel 660p, the first NVMe SSD with 4-bit per cell flash. When released the $199.99 660p broke new ground for NVMe pricing. The new MP510 gets within $36 of the 660p lessening its cost advantage and overall impact on this market.
The Adata SX8200 and HP EX920 are two of the drives to beat and have been very popular throughout 2018. The new MyDigitalSSD BPX Pro quickly challenged the two popular drives from the summer but is nearly identical to the MP510.
It's impossible to leave industry giants Samsung and SanDisk/Western Digital off our list. The Samsung 970 EVO and Samsung 970 Pro both make an appearance. The SanDisk Extreme Pro 3D NVMe also appears on the performance charts. The drive is identical to the Western Digital Black 3D NVMe.
Sequential Read Performance
The Phison PS5012-E12 has an issue with our sequential read test that comes across as a strong dip in the queue depth (QD) 2 and 4 results. This is likely something in the power management programming. The SanDisk Extreme Pro has the same issue but it must more severe.
Later in the review, we will show the MP510, one of the E12 drives, delivering strong sequential read performance that makes us dismiss this result as an oddity in testing.
Sequential Write Performance
The BPX Pro's orange result covers most of the Corsair MP510 line in the sequential write test. The two drives deliver nearly identical write performance that surpasses every other drive in the charts today, including the Samsung 970 Pro in this series using burst data.
Sustained Sequential Write Performance
The SLC cache for the MP510 is very strong. Corsair uses a dynamic cache that shrinks as you add data to the flash. With the drives nearly empty, we write to the entire user area of the flash with 128KB sequential data.
We did spot a slight difference in our results between the BPX Pro and the MP510. You can see where both drives drop out of SLC mode, there is a large dip in write performance. The MP510 dropped much lower than the BPX Pro for a short time before both leveled off around 1,400 MB/s.
Random Read Performance
The one area we would like to see the MP510 perform a little better is in the random read category. This may be one of the areas addressed with the new firmware. We don't have confirmation on that, so at this point, we're just hoping to see this addressed. Corsair, as well as MyDigitalSSD, chose Toshiba's 64L TLC memory using the Toggle 2 interface. The new 96L memory uses Toggle 3 and with the new revision comes a faster bus speed that should increase random read performance at QD1.
Random Write Performance
The 1TB MP510 nearly reaches single manager peak performance at QD2 in the random write test. The result is very close to 100,000 IOPS. The drive delivers nearly 60,000 IOPS at QD1, and along with the BPX Pro, the two drives take the top spot on our line chart.
70% Read Sequential Performance
The synthetic charts show weighted to the lowest queue depth and in the mixed workload chart that's QD2 because you can't read and write at the same time doing just one operation at a time. At QD2, the MP510 trails some of the other drives and appears in the middle of the chart. If you look at the QD4 performance, the middle bar on the second chart, the MP510 leads all of the other drives. That also carries over to QD8.
70% Read Random Performance
The 1TB MP510 is down in the random mixed workload test, and that will show up later on the next page in some of the real-world application tests. Let's get to the applications now that we see where the MP510 stands against the other popular drives in synthetic performance.
PRICING: You can find the product discussed for sale below. The prices listed are valid at the time of writing, but can change at any time. Click the link below to see real-time pricing for the best deal:
- We at TweakTown openly invite the companies who provide us with review samples / who are mentioned or discussed to express their opinion of our content. If any company representative wishes to respond, we will publish the response here.