Benchmarking the DELTA MAX SSD
Benchmarking SSDs are not an exact science as there is variability between runs, and different benchmarks may show different results depending on how they run their tests and how up-to-date the benchmarks are. However, by using enough real world and synthetic tests, it may be possible to get a good idea of the relative performance across all six tested drives. For benchmark results, the drives are generally listed in the following order on the charts:
- T-FORCE DELTA MAX SATA III 1TB SS
- T-FORCE Vulcan 500GB SATA III SSD
- Team Group MP33 1TB NVMe Gen 3 x4 PCIe SSD
- T-FORCE Liquid CARDEA 512GB PCIe Gen 3 x4 NVMe SSD
- T-FORCE CARDEA IOPS 1TB PCIe Gen 3 x4 NVMe SSD
- T-FORCE CARDEA Ceramic C440 1TB PCIe Gen 4 x4 NVMe SSD
We did not set up Windows on the Vulcan SSD, so not all of the benchmarks could be run on it. All of the drives will have their results summarized in multiple charts. Let’s start first with TeamGroup’s own S.M.A.R.T. utility to get some information on each SSD.
This TeamGroup S.M.A.R.T. utility is one that we recently discovered which gives more information in one place than many other disk tools.
First is the system and disk information for the DELTA MAX
S.M.A.R.T. also tests the drive with two sets of tests. First up, the DELTA MAX dual performance tests starting with MB/s.
Next the DELTA MAX performance dual test results given in IOPS.
We look at the performance tests in MB/s for the Vulcan SATA III SSD. Finally we look at the dual latency tests for the Vulcan SSD.Next we look at the performance tests in MB/s for the TeamGroup MP33 SSD. Next up, we look at the dual latency tests for the MP33 SSD.
Next the dual performance tests in MB/s for the T-FORCE CARDEA IOPS.
Finally, S.M.A.R.T. measures latency for the CARDEA IOPS.
Next, the performance tests in MB/s for the T-FORCE CARDEA Ceramic C440. Next, S.M.A.R.T. measures latency for the CARDEA Ceramic C440.S.M.A.R.T. clearly shows the PCIe CARDEA Ceramic C440 is the fastest SSD, followed by the CARDEA IOPS, then the CARDEA Liquid, then the TeamGroup MP33 SSD, and in last places, the SATA III Vulcan SSD and the DELTA MAX offer similar performance although they do not reach their stated Read/Write speeds of 560/510 MBps.
TeamGroup’s S.M.A.R.T. tool is a great place to start, so let’s see what other synthetic and real world tests show. Let’s begin with PCMark 10.
UL (formerly Futuremark) has been a developer and publisher of PC benchmark applications for nearly two decades. Although PCMark benches are synthetic tests, they provide a good measure of system performance. PCMark 10 was primarily developed for Windows 10 and it builds upon the PCMark 8 platform for a package of vendor-neutral home and office benchmarks.
Unfortunately, in comparison to PCMark 8, PCMark 10 is missing key elements including detailed storage testing, and it is not possible to test the Vulcan SATA SSD as an attached drive as with the older suite. We use both PCMark 10’s Express and Extended settings which are summarized after all of the screenshots are presented.
The Express benchmark suite is best suited for office tasks while the Extended benchmarks are for power users To properly compare the PCMark 10 scores, look at the detailed results that you want to compare. All of the SSD results are presented as screenshots. Open the images in separate tabs for easy individual test result comparisons.
First up is the T-FORCE DELTA MAX White SATA III Express score with 5045.
Next up is the CARDEA IOPS NVMe Express score with 5610.
First up is the T-FORCE DELTA MAX White SATA III Extended score with 3530.
Next up is the CARDEA IOPS Extended score with 3725.
Here is the summary of the five drives that were tested.
Let’s check out PCMark 8 which uses dedicated storage tests.
To see exactly where drive performance results differ, there is no better tool than SiSoft’s Sandra 2020. Sandra (the System ANalyser, Diagnostic and Reporting Assistant) is a complete information & diagnostic utility in one package. It is able to provide all of the information about your hardware, software, and other devices for diagnosis and for benchmarking.
AIDA64 is the successor to Everest and it is an important industry tool for benchmarkers. AIDA64’s benchmark code is written in Assembly language, and they are well-optimized for AMD, Intel and VIA processors by utilizing the appropriate instruction set extensions. We use the Engineer’s version of AIDA64 courtesy of FinalWire. AIDA64 is free to to try and use for 30 days.
- The Linear Read test measure sequential performance by reading or writing all sectors without skipping any. It’s a linear view of the drives overall performance from its beginning to end.
- The Random Read test measures the random performance by reading variable-sized data blocks at random locations on the drive and they are combination of both speed and access times as its position changes before each new operation.
- The Buffered Read test measures the drive caching.
- The Access time tests are designed to measure the data access performance by reading 0.5 KB data blocks at random drive locations
The individual benchmarks take much longer and they are more accurate and they are presented below without comment. Here is the summary chart comparing our six tested drives where higher is better except for the Average Read Access where lower is better.
SPECworkstation3 (3.0.4) Disk Benchmarks
All the SPECworkstation3 benchmarks are based on professional applications, most of which are in the CAD/CAM or media and entertainment fields. All of these benchmarks are free except for vendors of computer-related products and/or services. The most comprehensive workstation benchmark is SPECworkstation3. It’s a free-standing benchmark which does not require ancillary software. It measures GPU, CPU, storage and all other major aspects of workstation performance based on actual applications and representative workloads. SPECworkstation Disk benchmarks are perhaps more demanding than the 3DMark tests and only WPCstorage was performed. It was not possible to run it on the Vulcan SSD since there is no operating system installed on it.
Anvil’s Storage Utilities
Anvil’s Storage Utilities is a tool designed to benchmark and evaluate the Read and Write performance of SSDs and HDDs. It gives overall bandwidth as well separate Read and Write scores, the response times, and IOPS capabilities.
CrystalDiskMark is a HDD benchmark utility for your drives that measure sequential and random read/write speeds. Here are some key features of “CrystalDiskMark”:
- Measure sequential reads/writes speed
- Measure random 512KB, 4KB, 4KB (Queue Depth=32) reads/writes speed
- Results given in IOPS or MB/s
TxBENCH is similar to CrystalDiskMark but with additional features including secure erase. According to the website, “It not only measures the performance of storage easily but also performs detailed speed measurements based on specified access patterns and long-period speed measurements. It also allows you to see each drive’s supported features, enabled features, and S.M.A.R.T. information.”
This free standalone synthetic test is old and it doesn’t represent real world performance but it does test some important drive metrics. There is also a pay-for HD Tune Pro which is up-to-date and offers more functionality. We tried the Pro trial recently just to make sure the free version is still relevant. HD Tune has the following functions, and it measures the performance of:
- Transfer Rate
- Access Time
- CPU Usage
- Burst Rate
- Random Access test
- Write benchmark
Hard Disk information includes partition information, supported features, firmware version, serial number, disk capacity, buffer size, transfer mode.
- Hard Disk Health
- S.M.A.R.T. Information (Self-Monitoring Analysis and Reporting Technology)
- Power On Time
- Error scan
- Temperature display
Next we benchmark using AS SSD.
AS SSD is designed for Solid State Drives (SSD). This tool contains synthetic and practice tests. The synthetic tests determine the sequential and random read and write performance of the SSD without using operating system caches. In Seq-test the program measures how long it takes to read and write a 1GB file.
In the 4K test, read and write performance for random 4K blocks are determined. The 4K-64-thrd test are similar to the 4K procedure except that the read and write operations on 64 threads are distributed as in the usual start of a program. For the copy test, two large ISO file folders are created, programs with many small files, and a games folder with small and large files. These three folders are copied by the OS copy command with the cache turned on. AS SSD gives an overall score after it runs the benchmarks.
Here are the T-FORCE CARDEA Ceramic C440 AS SSD results also showing MB/s next to IOPS and below it, copy speeds.
Again, the CARDEA Ceramic C440 stands out even though it’s bandwidth is limited by Intel’s latest enthusiast motherboards which brings it’s performance just ahead of the IOPS. The CARDEA Liquid has a slower Write speed, but it is much faster than the MP33 or the Vulcan SATA SSD which again edges out the DELTA MAX. Game copy speeds are much closer for all tested SSDs.
Next up, ATTO.
ATTO is a low level hardware benchmark for random access read/write storage devices including for SSDs and HDDs. HD Tach uses custom device drivers and low level Windows interfaces to determine the physical performance of the device.
The T-FORCE DELTA MAX SATA III ATTO results are first.
HD Tach is up next.
HD Tach is a low level hardware benchmark for random access read/write storage devices that was developed by Simpli Software. HD Tach uses custom device drivers and low level Windows interfaces to determine the physical performance of the device. It is no longer supported and needs to be run in compatibility mode for Windows 10.
We present the benchmarks first with the Quick benchmark (8MB zones) on the left and the Long benchmark (32MB zones) on the Right.
Here are the DELTA MAX SATA III HD Tach results with an average read of 273.5MB/s for the Quick bench and 272.8MB/s for the Long bench.
Here are the T-FORCE CARDEA IOPS HD Tach results with an average read of 1508.4MB/s for the Quick bench and 1470.1MB/s for the Long bench.
Finally, here are the T-FORCE CARDEA Ceramic C440 HD Tach results with an average read of 1600.1MB/s for the Quick bench and 1553.4MB/s for the Long bench. Here are the HD Tach Disk benches summarized in a chart comparing our five drives. For read speeds, higher is better but for access times, lower is better.
The CARDEA Ceramic C440 is the fastest in HD Tach, and the Cardea Liquid is next followed closely followed IOPS, then by the MP33 SSD while the SATA III SSDs take the last two places where the Vulcan SSD is faster than the DELTA MAX .
Next we look at game/level loading speeds.
The Game/Level Loading Timed Results
Game and game level loading time results are difficult to measure precisely but generally SSDs perform similarly with regard to game loading times and they significantly faster than any HDD. Even SSHDs require loading the same level or program over-and-over to get quicker.
We tested 3 games using the World of Warcraft and Battlefield 3 loading times again as measured precisely by PCMark 8’s storage test as well as 5 levels and overall loading times accurately by using the Final Fantasy XIV: Stormbringer benchmark.
Let’s start with the Stormbringer benchmark first with the DELTA MAX SATA III . Total Loading times are 16.483 seconds.
Let’s start with the Stormbringer benchmark first with the CARDEA IOPS SSD. Total Loading times are 15.288 seconds.
Next, the Stormbringer benchmark with the CARDEA Ceramic NVMe PCIe SSD and the total Loading times are 13.34 seconds. Below is the chart comparing the total loading times for each SSD plus each of its 5 levels. Lower times in seconds are better which means the game or level will load more quickly and you will not have to wait as long to begin or to get back into the game.
Here are the World of Warcraft and Battlefield 3 loading times again as measured precisely by PCMark 8’s storage test. Lower (faster setup in seconds) is better.
All of the SSDs load games quickly and there is very little difference between them although the CARDEA Ceramic C440 is the fastest followed by the CARDEA IOPS and Liquid SSDs, then the MP33 and finally by the SATA III SSDs.
Until developers start to target SSDs for game storage, or perhaps after the next generation of consoles move away from hard disk drives, we may see SSDs achieve the game loading performance they are capable of. However, faster is better when a gamer is waiting to get back into a game.
Lets look at file copy speeds next.
File copy speeds are important to gamers especially when they want to quickly transfer their game files from one location to another. First we copy a 22.8GB folder containing Alien Isolation from its Steam folder to a desktop folder which is something we do regularly when setting up Steam games on multiple PCs.
Pay careful attention to the charts (in green) that show the consistency and speed of file copies. They tend to show the ups and downs where each SSD runs out of cache and how long it takes to empty it and refill it.
22.8GB File Copy
The DELTA MAX took 2 minutes and 27.7 seconds. to copy 22.8GB
It took about 2 minutes and 22.5 seconds to copy the same Alien Isolation 22.6GB Steam folder from program files to the desktop using the Vulcan SSD.
The CARDEA IOPS took 25.6 seconds.
The CARDEA Ceramic C440 took 20.5 seconds or just 5 seconds faster than the IOPS.
Next we try something more time consuming when we copy multiple folders totaling 44.2GB from from Steam’s common files to a desktop folder.
44.2GB File Copy
The DELTA MAX took 7 minutes and 4.0 seconds. to copy a 44.2GB Steam folder from program files to the desktop.
Even when the CARDEA Ceramic C440 SSD is constrained severely by the Intel platform’s limited bandwidth, it still excels at copying compared to any of our other tested SSDs. The IOPS comes in second place well ahead of the CARDEA Liquid and MP33 PCIe SSDs which are in turn faster than either SATA III SSD.
Finally, let’s revisit game/level loading times plus all of our Summary charts and then reach our conclusion.