Today on my operating table is a cool flash drive for iPhone, iPad or iPod Touch from ADATA - i-Memory UE710. The main feature of this flash drive is the ability to connect it to iDevices via the Lightning port, and on the other side to a computer via a USB 3.0 (or 2.0) port. In this review I will try to describe its capabilities, as well as its strengths and weaknesses.

i-Memory Flash Drive comes in a small cardboard box with a plastic insert. The kit includes the flash drive itself and instructions in a variety of languages. Including in Russian. There is almost nothing written in the instructions, so we will figure it out ourselves.

The size is indicated on the front of the box (and on the flash drive itself). There are 3 options: 32, 64 and 128 gigabytes. In principle, it is logical that the line does not include 16 gigabytes (because it is not enough) and 256 gigabytes (because it would be very expensive).

On the reverse side are the device dimensions and system requirements. The flash drive is very light - only 16 grams. Fits in any pocket.

The top side of i-Memory is glossy black. Dust and fingerprints strive to leave their mark on the surface. There are also white and pink to choose from.

The reverse side in my version is gray. It can be seen that the body of the flash drive is tightly fastened from 4 parts.

Now comes the fun part. On the side of the flash drive there is a switch that has 3 positions:

• Left – USB available for connecting to a computer/TV
• Right – Lightning available for connecting to iPhone, iPad or iPod Touch 5, 6
• Central – both connectors inside the case. Convenient to carry.

Switching the position of the lever occurs with little effort and characteristic clicks. The positions are fixed very well, so the flash drive will not have a situation where you plug it into the port and the connector starts to crawl inside the case on its own. This is what annoys me about many regular USB flash drives.

Connect i-Memory to the device:

Program for interacting with iMemory flash drive

Unlike many Chinese counterparts, AData focuses on its built-in application, which is actively supported in the App Store. I'll explain it popularly. By purchasing a Noname flash drive on some Chinese site, you risk that it will only download photos from you. iOS is a very closed system and without jailbreak, even the official Apple accessory, the Camera Connection Kit, loses the lion's share of its functionality.

AData did a very (cleverly) smart thing - they released their own program in the App Store, which greatly expands the functionality of the drive. When you connect to the flash drive for the first time, you will be prompted to install the program. The next time, the system will prompt you to open the program. Very good and useful.

What you can do through the built-in program:

• transfer photos and videos from phone memory
• backup contacts
• create folders (that is, you can scatter files among them)
• copy, cut and paste files
• watch photos and videos, listen to music, view documents

I will dwell on the last point in more detail:

The sign from the official website shows the seriousness of AData to capture the market. At first I didn’t believe it, but then I started testing mp3, mkv, avi... The program easily coped with the task. Yes, you don’t even have to copy files anywhere, but simply view them from a flash drive.

Watch video:

Listening to music:

Screenshots from the application:

Files can be laid out on a flash drive as you please. In folder Conductor, you will see the entire file structure.

It would seem perfect! But not really. I identified only 2 shortcomings, and the first one is most likely Apple’s fault.

1. The program does not see music due to iOS limitations.

I quote the FAQ:

Apple does not allow content to be moved directly to/from iTunes from iPhones or other iOS devices. However, most songs can be transferred to i-Memory UE710 using a Mac or PC and then listened to. All new songs from iTunes can be played. A small number of older songs are still protected by the service's stricter DRM (Digital Rights Management) rules and can only be played properly on iTunes.

My program did not see either licensed purchased music or pirated mp3s uploaded to the phone. So you can only download music from a PC or Mac and listen to it from a flash drive, but not import it in any way.

2. Sometimes viewing files with the built-in program does not satisfy me, but opening the file in another program is not possible.

a) I want to watch a movie, but the file contains 3 audio tracks. The built-in program, with all due respect, takes only the first track.

b) I look at the PHP file, and the program says that this is an unsupported type. OK. But how can I open it with another program?

Fortunately, small files can be sent by mail to yourself, uploaded to Dropbox, or sent by message. But you can't do that with films.

Maybe the developers will add this feature. I already wrote to them about this - it turned out that AData has Russian support. I also gave them a couple of minor but unpleasant bugs with the Cyrillic alphabet. Surely they will fix such a small thing.

Important. About formatting. Out of the box, i-Memory is formatted in OS X Extended format. I recommend immediately reformatting it to FAT 32. The fact is that your phone and tablet absolutely do not care what format the flash drive is in, but some TVs do not understand the OS X format. I came across this when I recorded a bunch of films for a trip to my parents... So feel free to format it!

A-Data managed to surprise me with its application. Even jailbreak is not necessary for almost full work with additional memory. And the quality of i-Memory UE710 is at a decent level. I recommend it to expand the memory of your favorite Apple gadgets. Nice stuff!

P.S. I took the flash drive as a gift for my younger sister on the condition that I would test it for at least two weeks for review. She dreamed about it, because she loves to take pictures and shoot videos, and her iPhone has only 16 gigabytes of memory. I think it's a wonderful gift. Don't you find it?)

from correspondence with my sister :)

We continue to explore the critical low-level characteristics of high-speed DDR2 modules with our all-in-one benchmark suite. Today we will look at modules from the Taiwanese manufacturer A-DATA, the Vitesta series, designed for a frequency of 800 MHz (in DDR2 mode). Information about the module manufacturer

Module manufacturer: A-DATA Technology Co., Ltd.
Module chip manufacturer: Elpida Memory, Inc.
Module manufacturer website: Chip manufacturer website: Module appearance

Photo of the memory module

Photo of memory chip

It was interesting to discover that after removing the radiator (which was a fairly easy and painless procedure for the modules), the appearance of the modules remained quite decent and complete. In particular, both on the radiator of the modules and on the modules themselves there are stickers with Part Number and serial numbers, and in the second case even with the company logo, which the average user is unlikely to be able to see :). Part Number of module and chip

Module Part Number

What’s nice is that the module numbers on the stickers on the outside and inside match (the serial numbers, however, no longer match). However, there is no guide on decoding the Part Number of Vitesta DDR2 memory modules on the manufacturer’s website. On the module description page, only the main technical characteristics are indicated: the capacity of each module is 256 or (in our case) 512 MB, the modules are based on 32Mx8 microcircuits, operate with a delay value of CAS# = 5 and a standard supply voltage of 1.85±0.1V.

Part Number of the chip

Description of technical characteristics (data sheet) of 256-Mbit DDR2 Elpida memory chips:

Note that the marking of the microcircuits of this module does not somewhat correspond to the official specification given in the data sheet of memory chips. Namely, the marking begins with a single letter “E” instead of the expected letter combination “EDE”, in addition, there is no package type code (“SE” = FBGA packaging). However, we have seen a similar approach to marking Elpida chips before (for example, on Kingston DDR2 module chips), so this is more the rule than the exception. SPD module chip data

Description of the general SPD standard:

Description of the specific SPD standard for DDR2:

Parameter	Byte	Meaning	Decoding
Fundamental memory type	2	08h	DDR2 SDRAM
Total number of module line address lines	3	0Dh	13 (RA0-RA12)
Total number of module column address lines	4	0Ah	10 (CA0-CA9)
Total number of physical banks of the memory module	5	61h	2 physical banks
External memory module data bus	6	40h	64 bit
Supply voltage level	8	05h	SSTL 1.8V
Minimum duration of the clock period (t CK) at maximum delay CAS# (CL X)	9	25h	2.50 ns (400.0 MHz)
Module configuration type	11	00h	Non-Parity, Non-ECC
Type and method of data regeneration	12	82h	7.8125 ms 0.5x reduced self-regeneration
Width of the external data bus interface (type of organization) of the memory chips used	13	08h	x8
Width of the external data bus interface (type of organization) of the ECC module memory chips used	14	00h	Undefined
Duration of transmitted packets (BL)	16	0Ch	BL = 4.8
Number of logical banks of each chip in the module	17	04h	4
Supported delay lengths CAS# (CL)	18	38h	CL = 5, 4, 3
Minimum clock period duration (t CK) with reduced CAS# delay (CL X-1)	23	30h	3.00 ns (333.3 MHz)
Minimum clock period duration (t CK) with reduced CAS# delay (CL X-2)	25	3Dh	3.75 ns (266.7 MHz)
Minimum time for recharging data in a row (t RP)	27	32h	12.5 ns 5,CL=5 ~4.2, CL = 4 ~3.3, CL = 3
Minimum delay between activation of adjacent rows (t RRD)	28	1Eh	7.5 ns 3,CL=5 2.5, CL = 4 2,CL=3
Minimum delay between RAS# and CAS# (t RCD)	29	32h	12.5 ns 5,CL=5 ~4.2, CL = 4 ~3.3, CL = 3
Minimum pulse duration of the RAS# signal (t RAS)	30	2Dh	45.0 ns 18, CL = 5 15, CL = 4 12,CL=3
Capacity of one physical memory bank	31	40h	256 MB
Recovery period after recording (t WR)	36	3Ch	15.0 ns 6,CL=5 5,CL=4 4,CL=3
Internal delay between WRITE and READ commands (t WTR)	37	1Eh	7.5 ns 3,CL=5 2.5, CL = 4 2,CL=3
Internal delay between READ and PRECHARGE commands (t RTP)	38	1Eh	7.5 ns 3,CL=5 2.5, CL = 4 2,CL=3
Minimum row cycle time (t RC)	41, 40	39h, 30h	57.5 ns 23, CL = 5 ~19.2, CL = 4 ~15.3, CL = 3
Period between self-regeneration commands (t RFC)	42, 40	4Bh, 30h	75.0 ns 30, CL = 5 25, CL = 4 20, CL = 3
Maximum duration of the clock period (t CK max)	43	80h	8.0 ns
SPD revision number	62	12h	Revision 1.2
Checksum byte 0-62	63	B9h	185 (correct)
JEDEC Manufacturer Identification Code	64-71	7Fh, 7Fh, 7Fh, 7Fh, CBh	A-DATA Technology
Module Part Number	73-90	—	Undefined
Module production date	93-94	00h, 00h	Undefined
Module serial number	95-98	00h, 00h, 00h, 00h	Undefined

The SPD content looks almost standard. The modules support all three possible values ​​of CAS# signal delay 5, 4 and 3. The maximum value corresponds to a clock period of 2.50 ns (frequency 400 MHz, i.e. nominal DDR2-800 mode) and a completely normal timing scheme 5-5-5- 18. Reduced latency CAS# (CL X-1 = 4) is prescribed for use in DDR2-667 mode (clock period 3.00 ns, frequency 333.3 MHz). Unfortunately, it is not possible to apply integer timing values ​​for this case; rounding to the nearest tenth gives the scheme 4-4.2-4.2-15, which will most likely be perceived by motherboard BIOSes as 4-5-5-15 (rounding to larger side for reasons of greater stability). The last, twice reduced value of t CL (CL X-2 = 3) corresponds to the DDR2-533 mode (cycle time 3.75 ns, frequency 266.7 MHz). The timing scheme for this case also turns out to be non-integer 4-3.3-3.3-12 (i.e. in reality 4-4-4-12). The manufacturer code indicated in the SPD, which is pleasant, corresponds to reality, however, there is no data on the date of manufacture, Part Number and serial number - and this does not make a very good impression. Test bench and software configurations

Test bench No. 1

• Processor: Intel Pentium 4 560, 3.6 GHz (Prescott rev. E0 core, 1 MB L2)
• Chipset: Intel 955X, FSB frequency 200 MHz
• Motherboard: Gigabyte 8I955X Pro, BIOS version F5 from 07/05/2005
• Memory: 2x512 MB A-DATA DDR2-800, single/dual channel
• Video: Leadtek PX350 TDH, NVIDIA PCX5900
• HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
• Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.2.1.1003, DirectX 9.0c

Test results

Performance tests

For a number of reasons, among the motherboards that tested A-DATA DDR2-800 modules, there was only one model: Gigabyte 8I955X Pro. Note that this board is one of the few that supports such fast memory as DDR2-800. As in our previous study, we carried out tests in both the usual dual-channel and single-channel modes in order to show the potential of DDR2-800 modules (in particular, bandwidth) in its “pure form”.

Parameter	Stand 1
	Dual channel mode	Single channel mode
Timings	5-5-5-15	5-5-5-15
Average read bandwidth, MB/s	5799	5770
Average bandwidth per write, MB/s	2456	2393
Max. Read memory bandwidth, MB/s	6457	6333
Max. Write bandwidth, MB/s	4279	4279
	44.0	44.0
	51.7	51.7
	93.7	93.9
	113.2	113.4
Minimum latency of pseudo-random access, ns (no hardware prefetch)	68.4	68.4
Maximum latency of pseudo-random access, ns (no hardware prefetch)	87.9	87.9
Minimum random access latency *, ns (no hardware prefetch)	94.1	94.3
Maximum random access latency *, ns (no hardware prefetch)	114.2	114.3

* block size 16 MB

The default timing scheme 5-5-5-15 (Memory Timings: “by SPD”) is slightly different from the scheme specified in the SPD (5-5-5-18). True, we can safely turn a blind eye to this difference, because, as we saw in the next series of tests, the modules in question are absolutely insensitive to the t RAS value specified in the chipset configuration registers, like most other DDR2 modules.

The speed indicators (BRP) of modules in dual-channel and single-channel modes are slightly different - of course, in favor of the dual-channel mode. The largest difference (albeit only 2%) can be seen in the maximum real memory bandwidth test for reading (6457 versus 6333 MB/s). The difference is small, however, in the previous study of Corsair XMS2-8000UL modules, almost no such difference was observed at all. It is possible, of course, that this is due to the use of different processors (Pentium 4 560 and 670) it is very likely that the larger L2 cache of the Pentium 4 670 processor is able to hide differences in memory bandwidth to a greater extent. However, the potential of the considered DDR2-800 “in its pure form” (that is, due to the real throughput of a single channel) also turns out to be quite good.

In addition, one cannot help but rejoice at the very low (noticeably lower, compared to the same Corsair XMS2-8000UL) latency values ​​when accessing memory, even with a standard timing scheme. However, this cannot be directly considered an advantage of these modules over the others; after all, the tests used different models of motherboards (and, most importantly, different BIOS versions) and different processors (the influence of this factor is much less likely, but still However, it should not be excluded either). Therefore, the final answer to the question of whether the memory modules under consideration are actually characterized by lower latencies requires additional research.

Stability tests

The timing values, with the exception of t CL, were varied on the fly thanks to the ability built into the RMMA test package to dynamically change the memory subsystem settings supported by the chipset. The stability of the memory subsystem was determined using the RightMark Memory Stability Test auxiliary utility, which is part of the RMMA test package.

To achieve minimum timings, we set the modules' supply voltage to a slightly higher 2.2V. Naturally, the experiment could have been carried out at a standard (for these modules) voltage of 1.85V, but, firstly, the result would hardly have been as indicative, and secondly, it could not be directly compared with the result achieved for Corsair XMS2-8000UL modules.

Minimum timing values ​​that allow you to set the memory modules in question in DDR2-800 mode without losing stability 4-4-4 (attempts to further reduce the t RP and/or t RCD values ​​led to an immediate system freeze). Of course, this clearly falls short of the previous records set by Corsair modules (4-3-3 for the XMS2-8000UL and even 4-3-2 for the XMS2-5400UL). At the same time, you will agree that the ability to set a timing scheme that is more typical for DDR2-533 (standard) and DDR2-667 (in fact) modules looks very good in the DDR2-800 high-speed mode.

Parameter	Stand 1
	Dual channel mode	Single channel mode
Timings	4-4-4 (2.2V)	4-4-4 (2.2V)
Average read bandwidth, MB/s	5841	5825
Average bandwidth per write, MB/s	2465	2421
Max. Read memory bandwidth, MB/s	6477	6367
Max. Write bandwidth, MB/s	4279	4279
	43.7	43.8
	50.9	51.1
	88.6	89.0
	107.9	107.7
Minimum latency of pseudo-random access, ns (no hardware prefetch)	67.9	68.0
Maximum latency of pseudo-random access, ns (no hardware prefetch)	87.4	87.8
Minimum random access latency *, ns (no hardware prefetch)	89.0	89.2
Maximum random access latency *, ns (no hardware prefetch)	109.0	109.0

* block size 16 MB

“Overclocking by timings” brought quite predictable changes to the test results: read bandwidth increased slightly (very slightly), random access latency decreased slightly. The gap between the performance of the dual-channel mode and that of the single-channel mode has narrowed somewhat (the maximum difference in the maximum real read bandwidth is now only 1.7%), which, in general, is also quite natural. Results

The tested A-DATA DDR2-800 memory modules of the Vitesta series showed themselves to be high-speed modules that can almost fully reveal the potential of this type of memory (tests in single-channel mode), and also have very low latencies (at least in the conditions of our experiment ). The overclocking potential in terms of timings of these modules is also quite good: with a supply voltage of 2.2V (typical for high-end “enthusiast” class modules), the modules operate stably with timings of 4-4-4 , more typical for entry-level DDR2-533 modules and high-end DDR2-667 modules. As in the case of our previous studies, it is too early to judge the compatibility of DDR2-800 memory modules with various motherboards, because The motherboards that actually support such high-speed memory operating modes can still be counted on the fingers of one hand.

02 July 2010, Odnokrylov Vladimir 0

DDR3 RAM, which has occupied new platforms from Intel and AMD, is acquiring new frequencies. We have already reviewed Kingston's Hyper-X speed kits. Today we will turn to the memory strips produced by A-DATA - the AX3U1600GC4G9-2G set with a capacity of 8 GB and the set AX3U2000XB2G9-2X with a capacity of 4 GB.

Introduction

I continue the line of reviews dedicated to today’s most modern DDR3 RAM; this time we want to talk about the brand that is not yet the most popular in Russia and its neighboring countries - . Strictly speaking, the products of this company are not limited to desktop memory; for example, flash drives with this inscription may also have caught your eye more than once.

But it was the memory modules that fell into our tenacious hands, and even from the lines Gaming Series(G Series) and Extreme(X Series). But more on this a little further, but for now let’s get acquainted with the subjects.

Technical characteristics of A-DATA DDR3 1600G and A-DATA DDR3 2000X

Well, let's start looking at series sets DDR3 1600G And DDR3 2000X from company . The technical characteristics of the modules are as follows:

Marking	AX3U1600GC4G9-2G	AX3U2000XB2G9-2X
Volume	2x4 GB	2x2 GB
Microcircuits		16 DDR3 SDRAM chips in FBGA form factor
Design	240-pin DIMMs	240-pin DIMMs
Nutrition	1.55 - 1.75V	1.6 - 2.0V
Data transfer rate	1333/1600(o.c.) MHz	1333/2000 (os) MHz
Delay mode	9-9-9-24/9-9-9-27	9-9-9-24

We received two plastic boxes with cardboard inserts, painted in dark colors.

As it turned out, the letters at the end of the series marking distinguish the relationship of the modules to a specific line of memory, G means a game series, and X- extreme series.

Opening the box showed that the trims of different series also have different radiator structures. The arrangement of microchips in each module is double-sided - hence the radiators.

A-DATA DDR3 1600G and A-DATA DDR3 2000X | Test bench

We will test RAM on the following configuration:

CPU	AMD Phenom II X6 1055T
Motherboard	ASRock 890GX Extreme3
Video card	Integrated ATI Radeon HD 4290 (368 MB, Catalyst 10.4)
power unit	Tuniq Ensemble 1200 W
HDD	WD7501AALS 750GB
OS

Testing A-DATA DDR3 1600G and DDR3 2000X

The manufacturer has special overclocking profiles built into the tested modules of the "extreme" series. We didn’t have to try them, that is, get these frequencies in the test configuration - we remind you that the purpose of the test is to check the relevance of high-speed memory modules for motherboards that do not support frequencies above 1800 MHz. Briefly, the information about the modules obtained using :

A-DATA AX3U1600GC4G9-2G A-DATA AX3U2000XB2G9-2X

So during the testing process, we simply compared modules of different series with each other. The tests were carried out in dual-channel unganged mode, the modes were set automatically by the motherboard. Memory operating modes with timings are shown in the table below.

Marking	AX3U1600GC4G9-2G	AX3U2000XB2G9-2X
1333 MHz	667 MHz, 9-9-9-24	666 MHz, 9-9-9-24
1600 MHz	795 MHz, 9-9-9-27	800 MHz, 9-9-9-24
1333 MHz with AMD Turbo Core technology	638 MHz, 7-7-7-20	636 MHz, 7-7-7-20

The results of memory speed tests are given below.

When the question of building a system based on the LGA1366 platform is raised, it is likely that kits designed for an operating frequency of 1600-1800 MHz will be used as memory. This is the best case scenario. At worst, when a finished PC is purchased, DDR3-1333 modules with a total capacity of 3-6 GB will modestly “take shelter” in the system unit. Of course, with a three-channel memory controller in Intel Core i7-9xx processors, the frequency of the processors is not so critical, but what Russian doesn’t like driving fast? Taking into account a certain safety margin that the manufacturer initially included in its products, it is quite possible to achieve higher frequencies than those for which the memory is officially designed. We will try to find out whether this is true using the example of a six-gigabyte DDR3-1333 memory kit produced by A-Data.

A-Data DDR3-1333G

Although A-Data's products were not widespread in our market, some memory modules could sometimes please enthusiasts and overclockers with their potential. The strips are supplied in a sealed blister with an insert label, which has nothing interesting on it.

A-Data DDR3-1333G modules (the kit has the article number AD31333G002GU3K, and each module has the AD31333G002GMU) belong to the series for gamers, are made on a green PCB and are equipped with standard black aluminum radiators, which are attached to the chips using thermal Velcro.

A sticker is glued to each half of the radiator - one of them contains information about the operating frequency, timings and voltage, and the second contains a bar code and memory article number. And if the frequency of 1333 MHz and delays like 8-8-8-24 are unlikely to attract attention, then the voltage of 1.65 ~ 1.85 V can be a little confusing. But the values ​​in this limit are quite safe - the main thing is that the delta between the supply voltage of the memory and the memory controller is about 0.5 V.

The SPD of memory sticks contains only standard timings for frequencies from 832 to 1333 MHz (according to the Everest program): 5-5-5-15 for 832 MHz, 6-6-6-18 for 1000 MHz, 7-7-7- 21 for 1166 MHz and 8-8-8-24 for 1333 MHz. There is no XMP profile, and it is not needed, since everything is already specified without it. Only if necessary will you need to set the memory supply voltage to 1.65 V instead of the standard 1.5 V.

In addition to four sets of delays and frequencies, the MemSet utility can detect a fifth one, which at a frequency of 1500 MHz allows you to set timings like 9-9-9-27.

Test configuration and overclocking technique

The memory was overclocked on the following configuration:

• Processor: Intel Core i7-965 (3.2 GHz, C0);
• Motherboard: DFI X58-T3H6 (Intel X58);
• Video card: ASUS EN8800GS TOP (GeForce 8800 GS 384MB);
• Cooler: Noctua NH-U12P;
• Hard drive: Samsung SP2504C (250 GB, SATA2);
• Power supply: Seasonic SS-600HM (600 W).

Testing was carried out in Windows Vista Home Premium x64 SP2, using four running copies of the LinX program using 1024 MB of memory to check system stability.

The ratio of the frequency of the clock generator, the multiplier on the memory and the processor in the BIOS Setup of the motherboard was selected individually, but more often the CPU multiplier was x23 or x21, and the Bclk frequency was in the range of 133-162 MHz. The QPI bus throughput was 4800 MT/s. The voltage on the memory controller was set at 1.36 V, since no positive effect was observed at a higher level. The memory voltage was 1.65 V. The remaining BIOS settings did not affect the overclocking level and were set to Auto.

The overclocking potential was determined for three sets of timings that are currently relevant for DDR3 memory: 7-7-7-21, 8-8-8-24 and 9-9-9-27 with Command Rate 1T. Minor delays remained at Auto.

Overclocking results

The potential of the A-Data DDR3-1333G memory surprisingly turned out to be quite good, and with timings of 7-7-7-21 we managed to reach 1510 MHz. The clock generator frequency (Bclk) in this mode was 151 MHz, the processor multiplier was x23, and the memory multiplier was x10.

Having set the standard delays for these bars to 8-8-8-24, the maximum frequency turned out to be 1620 MHz, Bclk was equal to 162 MHz, the multiplier on the processor and memory corresponded to x23 and x10.

Less aggressive timings - 9-9-9-27 - did not affect the result in any way, even with a change in Bclk, and the frequency remained at 1620 MHz.

conclusions

You can save on memory when purchasing a new set or increase the performance of a system with a Core i7 by increasing the frequency of an existing set without any problems, as our short material confirms. And although a memory kit with not the best potential was used, its low cost will allow you to spend an extra 50-70 dollars on an additional hard drive or “super cooler”, which will be enough to cool the junior representative of the Core i7-900 series, fortunately, the capabilities of the A-Data AD31333G002GU3K kit allow you to achieve both minimum 4 GHz when overclocking the processor. High timings at 1600 MHz