How can I know what speed card to get for my camera?

Deciding what card to buy for a specific camera in order to get the fastest performance of which the camera is capable without spending more than necessary on a card that is even faster than the camera is capable of taking advantage is a daunting task. Not only can it be very confusing when comparing speeds of various cards, but learning how fast a write speed your camera is capable of can be near impossible. Manufacturers rarely publish the maximum speeds at which their cameras can write to a sufficiently fast memory card. And many times faster, larger, newer cards can be cheaper than the older cards their predecessors replaced!

The current methods of rating various card speeds is, to put it mildly, A MESS. While the majority of cameras currently produced use some form of SD card (SD, SDHC, SDXC, and SDIO), many of the top professional models use CF cards. Even cameras that have slots for both types of cards often perform better in terms of speed when using a top rated CF card.

Part of the confusion among CF and older SD cards is the difference between read speeds, which is what most 133x, 600x, or 1000x ratings are based on and write speeds, which are what most photographers are concerned with since it affects the camera’s performance when shooting large numbers of frames in quick succession. The newer Video Performance Guide (VPG) rating is based upon the cards minimum continuous write speed, expressed in MB/sec. Unfortunately the only two tiers we see seem to see in the current marketplace are VPG-20 and VPG-65. A few VPG-45 rated cards (the speed needed to perform smoothly with 1080P HD video) surfaced at one time, but most of those same cards now carry the VPG-65 rating needed to insure smooth performance with 4K video. And with many newer cameras now capable of writing over 100 MB/s with the fastest cards, knowing two cards are capable of 65MB/s doesn’t go a long way toward determining which is best for a particular camera if you’re more worried about how fast your camera’s buffer clears shooting stills in burst mode than you are meeting the minimum requirements for HD or 4K video.

With SD cards, the capacity of the card is the primary determination as to whether it is an SD, SDHC, or SDXC card. Base transfer speeds are required to be higher for SDHC cards than SD cards using base speeds, and also for SDXC cards compared to SDHC cards at base speeds. But there is nothing preventing a manufacturer from producing a lower capacity SDHC (or even SD) card that is capable of faster speeds than the required base transfer speed of an SDXC card. With SDHC and SDXC cards transfer speeds are rated by the UHS-I or UHS-II specification, but write speeds may still be slower. Standards for SD cards designated as belonging to a certain Speed Class Ratings were developed to address the need for cards to be rated at particular write speeds as well as read speeds. But there can still be much variation between different cards of the same speed class rating as well as different speeds obtained from the same card depending on such things as the frequency of soft errors, file fragmentation, and whether a single large file or multiple smaller files are being written to the card. Also note that an UHS-I rating is not the same as an UHS Class 1 (U1) rating!

Several years ago one go-to guide for Canon or Nikon shooters would have been Rob Galbraith’s Digital Photography Insights. But the information there has become dated. The most recent Canon cameras tested were the Canon 5D mkIII and 1D mkIV (No 1D X, which released just after the 5D3). There were many more cameras that had test results listed at one time, but many of those cameras don’t even appear on the drop down list any more and some of the cameras still in the drop down list no longer display any test data. (The page linked above does contain some good information regarding how memory cards work and how to identify a particular card). To the best of my knowledge there is no comprehensive database that tests a plethora of cards in each of many camera models currently on the market. I would love to be shown such a database that proves I am misinformed at this point! Camera Memory Speed does have a limited list of newer models and more recent cards.

So what can a camera owner do to find what the most efficient memory card is for a particular camera? One that is capable of the camera’s top speed without being overkill? Short of buying a bunch of cards and testing them yourself, there are places to look to find some help.

I always start by reading the camera’s manual. Even if the camera’s write speed isn’t explicitly spelled out, there are often clues. For instance, on page 61 and again on page 270 of the Canon 7D Instruction Manual (updated edition for firmware v.2 or later) we read the following:

“Figures in parentheses apply to an UDMA, 128GB card based on Canon’s testing standards.”

So we should probably consider an Ultra DMA (UDMA) compliant CF card.

My experience is that an internet search (google, bing, etc.) can also usually turn up something. Search google for “Canon 7D UDMA card speeds”, and one of the top results is this link. A member has posted results of a test he conducted with his 7D and various CF cards rated at different speeds. Substitute ‘Nikon D700’ for ‘Canon 7D’ and you get similar results for that camera. As with most google searches, the more you know before you start, the easier it is to find what you need. Include ‘UDMA-7″ in the 7D search and the top result leads you to this discussion which makes it fairly clear that an UDMA-7 card is no faster in the 7D, even after updating the firmware to version 2, than an UDMA-6 card. So for the 7D, an optimal card would be a UDMA-6 compliant card.

Say you have a Canon T5i. What can we learn about it? Page 89 of the manual indicates a UHS-I compatible 8GB card was used to generate the best burst performance. Note that the burst performance increases with the UHS-I card when shooting JPEGs but stays at 8 frames when shooting RAW. Canon Europe’s support site says a large capacity Class 6 or higher SD should be used when shooting video. So in the case of the T5i, the question would be whether you plan to shoot JPEG, RAW, or video most of the time. How you answer that would determine whether the UHS-I Class 10 card would be overkill or not.

As to how various cards of the same rating compare, reviews such as this one at tom’s HARDWARE are fairly easy to find and show that not all Class 10 cards will out perform all Class 6 cards. Once you’ve narrowed it down, you can always check the reviews for a particular card at These reviews for the SanDisk Extreme 32 GB SDHC Class 10 UHS-1 Flash Memory Card 80MB/s (SDSDXS-032G-X46) are fairly typical. There are usually several reviewers who will post results of their test with a certain card. Some reviewers will demonstrate more knowledge and objectivity than others. The review rating system at amazon is useful for pointing to the most helpful reviews.

Note that as the use of high speed, high capacity SD/SDHC/SDXC cards become more ubiquitous, the prices have trended downward. This is probably a result of both higher volumes and increased competition among suppliers. The fastest CF cards, though, can still get rather pricey. But for cameras with dual slots the fastest CF cards usually outperform the fastest SDXC cards (as of November 2015).

A SanDisk Extreme 32 GB SDHC Class 10 UHS-1 Flash Memory Card 80MB/s currently sells for under $30 US at A SanDisk 32GB Extreme Pro CF memory card – UDMA 90MB/s 600x currently sells at for $167 US. The SD card reads at 80MB/sec, the CF card reads at 90MB/sec and costs five times as much. A newer UDMA-7 Sandisk 32GB Extreme Pro CF Card rated at 160MB/sec is much cheaper at $53 US, but a newer version of the 32GB SDHC Class 10 UHS-1 Flash Memory Card 95MB/s rated at 95 MB/sec is only $25 US.

And as always, beware the deals on the top brands that seem too good to be true. The most expensive cards are also the ones most often counterfeited. The best way to protect yourself from buying a counterfeit card is to only buy from reputable sources that have a liberal return policy.

What with the different write speeds for cards rated the same based on their read speed and the fact that there have been issues with certain camera/card combinations based on the camera’s formatting method vs. the card’s firmware, the only real way to know for sure how a particular card will perform in a particular camera is to test each card in each camera and compare the performance. To do it scientifically would mean using multiple copies of each from different production runs.

As stated above, there is no such comprehensive database of which I am aware. In addition to the Rob Galbraith site listed above, Camera Memory Speed does virtually identical tests with a fairly limited number of the newest cameras and memory cards. The models listed at CMS pick up right about the same time the RG site stopped testing any newer cameras. One likely reason there is not such a database is that memory cards marketed and sold under the same name/model number are often not identical in their hardware or firmware from one batch to the next. And the various cards are introduced and then replaced so frequently that it would be a herculean task to keep up with the changes in the cards currently on the market, much less test them in all compatible cameras that are also constantly evolving with firmware updates and new model introductions!

The speed at which the camera’s buffer (not processor) can write to a card is affected by the way the card communicates with the camera just as it is affected by the camera’s hardware and firmware limitations. A “write” operation also includes the card’s controller sending certain responses back to the camera verifying the integrity of the data that was received.

In the end the marketplace often renders the question moot. Current cards, with higher capacities and speeds, are often cheaper to buy from sources that move a lot of cards in short time periods (i.e. amazon, B&H, or newegg), than older cards that are slower and smaller and have been languishing in smaller sellers inventories for months and even years! The difference in prices are usually based more on brand marketing than anything else. And all of the major brands (Lexar, SanDisk, Transcend, Kingston, etc.) get their components from the same handful of suppliers that actually manufacture the memory chips and controller chips. So do the off-brand names, but they usually buy the leftover components that may or may not have been good enough to pass the QC of the major brands’ buyers.

And if you buy a memory card via eBay, you never know if you’re getting a genuine branded card or a counterfeit. There are probably more fake SanDisk cards in circulation worldwide than genuine ones! So always buy memory cards from reputable dealers that will let you return a card if it isn’t up to spec.

sql server – Is there anyway I can speed up this large full-table query?

I have a query that selects from only one table and with one WHERE filter. However it takes a very long time to execute and even times out occasionally. This is likely because it is filtering about 4 million rows out from a table of 13 million rows (the other 9 million records are older than 2019), and it is returning all of the columns, of which there are 101 (a mix of datetime, varchar, and int columns). It has two indexes, a clustered one on its primary key interaction_id, and an unclustered index on interaction_date which is a datetime column that is the main filter. This is the query:

FROM (Sales).(dbo).(Interaction)
year(Interaction_date) >= 2019

Is there anything obvious I can do to improve this query’s performance by adding/tweaking indexes or tweaking the query itself? Before I go into an ETL processes or fight back on the group that needs this query (they are a hadoop sqooping team who insist they need to sqoop all of these records all the time with all of the columns), I want to see if I can make it easier on people by doing something on my end as the DBA.

The query plan by default ignores my non-clustered index on the interaction_date column and still does a full clustered index scan. So I then tried forcing it to use it by including WITH (INDEX(IX_Interaction_Interaction_Date)) in the select.

This forces it into the query plan startign with an index scan of the non-clustered index, with estimated rows 4 million but estimated rows to be read as all 13 million. Then after a short time it spends the rest of the execution on the key lookup of the primary clustered index.

But ultimately, it doesn’t seem to speed up the query at all.
Any thoughts on how I can handle this? Thanks.

dnd 5e – Can a Swords bard benefit from Blade Flourish’s speed increase again if they take the Attack action again?

The speed increase will stack because the effect does not have a duration

The Dungeon Master’s Guide errata (pdf link) added the “Combining Game Effects” section which states (emphasis mine):

Different game features can affect a target at the same time. But when two or more game features have the same name, only the effects of one of them—the most potent one—apply while the durations of the effects overlap. (…) Game features include spells, class features, feats, racial traits, monster abilities, and magic items. (…)

Since game features include class features one would think that we can conclude that two instances of Blade Flourish’s speed increase would not stack; but in this case, there is no overlapping duration. We can conclude this by looking to the Essentials Kit, which includes a “Effects with the Same Name Don’t Stack” section (page 3) stating (emphasis mine):

(…) But when two or more effects have the same proper name, only one of them applies while the durations of the effects overlap (a duration is a time span of 1 round or more). (…)

This defines what counts as a duration and in this case, the Blade Flourish’s speed increase lasts until the end of the current turn (less than 1 round), thus you can benefit from the speed increase multiple times if you took multiple Attack actions.

Best DNS configuration for have fast speed

Hi friends,

I’m configuring a cluster DNS, and i’m checking how can i have better speed in resolve the requests:

-I will create 3-4 DNS for have more redundance and security
-Each DNS server will be in different country, i’m thinking buy him in: OVH, Digital Ocean and Amazon. You recommend me buy servers in other providers for have more fast cluster DNS?
-I will install cPanel DNSOnly

Any recommendation in your experience for be more fast?.

Thank you very much friends.

How to speed up script compilation?

Is there a way to speed up script compile time when I change something in VS then click on Unity?

It takes 15 seconds. Is this normal?

How Can I speed up decrypting?

I have a recovery company retrieving my scammed bitcoins. They have moved them directly to my brand new wallet but the decryption process is very very slow. They say I should purchase more bitcoins and deposit them in my wallet to create history. Will that speed up the process?

PHP version preferable for speed


It’s better select default PHP version for clients at "native" or latest PHP version?.

My native version is PHP 7.3, and latest is… | Read the rest of

windows 10 – Speed up a sofware opening by using RamDrive as a cache

I’m a developer, the RAM is enough to keep the whole source code in memory but there is an impressive difference if I work on the files from a RamDrive and if I hope that Microsoft Windows will keep the files in memory.

I would expect Windows to use all the available RAM to cache the prefetch of the program that I opern more frequently but it doesn’t happen. How can I force windows to use the RAM and write changes to the SSD only 1 every x seconds?

linux networking – How to investigate why my download speed from my dedicated server is capped at around 550KB/s?

I have an arm7hf dedicated server with 250Mb/s bandwidth. It has only 2GB of RAM but it is OK for my usage. When downloading files from it my speed is capped at around 550KB/s. I have a connection of 40Mb/s.
My torrent are uploading and downloading at full speed (~25MB/s). I only have this problem when transferring files through HTTP or SCP.

If I use a download manager that uses multiple connections, I get the full speed.

I tried to tweak some settings but without luck:

net.core.wmem_max = 4194304
net.core.rmem_max = 12582912
net.ipv4.tcp_rmem = 87380 120478 4194304
net.ipv4.tcp_wmem = 87380 120478 4194304

I changed the mtu from 1500 to 9000 but I have the same issue:

eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 9000
    inet  netmask  broadcast
    inet6 fe80::2e08:1cff:fe00:4d0a  prefixlen 64  scopeid 0x20<link>
    ether 2c:08:1c:00:4d:0a  txqueuelen 1024  (Ethernet)
    RX packets 701185  bytes 54586611 (54.5 MB)
    RX errors 1  dropped 0  overruns 0  frame 0
    TX packets 1114280  bytes 1472461501 (1.4 GB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

Here is the output of tcpdump :

10:09:42.394396 IP (tos 0x0, ttl 54, id 63137, offset 0, flags (DF), proto TCP (6), length 52)
DEVICE_IP:PORT > SERVER_IP:PORT: Flags (.), cksum 0x21e7 (correct), ack 15993210, win 3839, options (nop,nop,TS val 2429423187 ecr 4488744), length 0
10:09:42.395128 IP (tos 0x0, ttl 64, id 28565, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0x090f), seq           16078482:16079930, ack 144, win 110, options (nop,nop,TS val 4488747 ecr 2429423180), length 1448
10:09:42.395460 IP (tos 0x0, ttl 54, id 63139, offset 0, flags (DF), proto TCP (6), length 52)
DEVICE_IP:PORT > SERVER_IP:PORT: Flags (.), cksum 0x0b46 (correct), ack 15999002, win 3839, options (nop,nop,TS val 2429423187 ecr 4488745), length 0
10:09:42.398319 IP (tos 0x0, ttl 54, id 63142, offset 0, flags (DF), proto TCP (6), length 52)
DEVICE_IP:PORT > SERVER_IP:PORT: Flags (.), cksum 0xe952 (correct), ack 16007690, win 3839, options (nop,nop,TS val 2429423190 ecr 4488745), length 0
10:09:42.401705 IP (tos 0x0, ttl 64, id 28586, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0x4364), seq 16108354:16109802, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423189), length 1448
10:09:42.402421 IP (tos 0x0, ttl 64, id 28588, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0xa68a), seq 16111250:16112698, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423189), length 1448
10:09:42.402919 IP (tos 0x0, ttl 64, id 28589, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0x4062), seq 16112698:16114146, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423189), length 1448
10:09:42.403403 IP (tos 0x0, ttl 64, id 28590, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0x3e67), seq     16114146:16115594, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423189), length 1448
10:09:42.403747 IP (tos 0x0, ttl 54, id 63149, offset 0, flags (DF), proto TCP (6), length 52)
DEVICE_IP:PORT > SERVER_IP:PORT: Flags (.), cksum 0x9e4b (correct), ack 16026890, win 3839, options (nop,nop,TS val 2429423196 ecr 4488746), length 0
10:09:42.404396 IP (tos 0x0, ttl 64, id 28593, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0xea65), seq 16118490:16119938, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423189), length 1448
10:09:42.404723 IP (tos 0x0, ttl 54, id 63150, offset 0, flags (DF), proto TCP (6), length 52)
DEVICE_IP:PORT > SERVER_IP:PORT: Flags (.), cksum 0x92fa (correct), ack 16029786, win 3839, options (nop,nop,TS val 2429423197 ecr 4488746), length 0
10:09:42.406288 IP (tos 0x0, ttl 64, id 28596, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0xc003), seq 16122298:16123746, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423198), length 1448
10:09:42.406937 IP (tos 0x0, ttl 64, id 28597, offset 0, flags (DF), proto TCP (6), length 1500)
SERVER_IP:PORT > DEVICE_IP:PORT: Flags (.), cksum 0x6fcb (incorrect -> 0x15a5), seq 16123746:16125194, ack 144, win 110, options (nop,nop,TS val 4488748 ecr 2429423198), length 1448

Do you see anything that explain my situation? If not how can I debug this issue?

google search console – Slow WordPress site speed with Bluehost

Probably the best thing to do is get yourself off what looks like a crowded under-resourced hosting server.

It takes almost 1 second for a 301 redirect and then another second before your page even starts to download.


enter image description here

Generally with hosting. you get what you pay for