• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

FRONTPAGE NVIDIA GeForce RTX 2080 and RTX 2080 Ti Review

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Robert17

Premium Member
Joined
Feb 1, 2011
Location
Republic of Texas
It will be a year before everything shakes out with this card, minimum. First gen of SSDs had no TRIM. Controllers and software had to catch up. This is dissimilarly similar in spirit.
 

UltraTaco

Member
Joined
Oct 28, 2017
Who cares for trim?
Taco ran trim free with Vista for many years(maybe3?)
It's the same for Ray tracing. Run a special demo, enjoy how it looks and then back to regular game with no fancy Rays..
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
Sorry I was lazy and didn't verify the posting, but you are correct it's 46%. Techpowerup also lists the 2080ti as only having a 28% improvement over the 1080ti in 4k.
Also the $999 MRSP is a pipe dream, I haven't seen a 2080ti listed for less than $1169.
Even if it was $999 it's still the smallest increase in performance for the largest increase in MRSP.
Sadly this lack of competition means the customer gets the short end of the stick.
Yep.. good t get the numbers right. :)

Like any launch, MSRP is a pipe dream. But be fair and compare apples to apples, regardless if the point remains the same. :)

It will be a year before everything shakes out with this card, minimum. First gen of SSDs had no TRIM. Controllers and software had to catch up. This is dissimilarly similar in spirit.
ssd's at that time had garbage collection, which yielded similar results to TRIM.
 

OC-NightHawk

Member
Joined
May 13, 2003
Who cares for trim?
Taco ran trim free with Vista for many years(maybe3?)
It's the same for Ray tracing. Run a special demo, enjoy how it looks and then back to regular game with no fancy Rays..

I could go on and on about Vist in its proper topic but yeah ray tracing right now seems like physx all over again.
 

ratbuddy

Member
Joined
Aug 24, 2007
Well, I'm in the 'darnit, Amazon' boat. Still no idea when my first-batch preorder Windforce card will ship. I also have an nvidia.com (10/22 batch) order, no idea when that will really ship either. I'll try to refrain from calling this a paper launch until the 27th, the supposed end of the delay window.
 

UltraTaco

Member
Joined
Oct 28, 2017
I am the Sr. Director of Marketing and Performance Analysis for the division at LSI (soon to be owned by Avago) that makes the SandForce Flash controllers. I have presented many times at the Flash Memory Summit in Santa Clara, California, on the topic of SSD operations for many years. I also have a blog on www.blog.lsi.com covering all aspects of Solid State Drives.

[…]

For the short explanation flash memory is organized in groups of pages where data can be written. Once a page is written, it cannot be rewritten until it is erased. But a page can only be erased within a group of typically 128 pages called a block. The complexity of writing data really starts to escalate in the case of random writes replacing previously written data. Random writes put the new data in previously erased pages elsewhere, peppering a block of valid data with “patches of invalid data.” In order to write new data to these patches, the whole block – all 128 pages – must be erased. But first all surrounding pages with valid data must be read and then rewritten to blank pages. The newly erased block of blank pages is then ready to save new data.

[…]

All NAND Flash-based SSDs use GC. Some use foreground GC and some use background or idle-time GC. The difference between them is covered in my blog http://blog.lsi.com/dont-let-ssds-throw-away-your-gold/ . In simple terms background garbage collection will increase write amplification (WA) and wear out the SSD sooner. Foreground GC is harder to achieve and I believe only the SandForce controller is able to do it today

[…]

TRIM is beneficial to all SSDs regardless of what kind of garbage collection is used. I talk about how TRIM came into existence and why it is necessary in my blog http://blog.lsi.com/did-you-know-hdds-do-not-have-a-del… . The TRIM command is sent by the OS to the SSD to identify what pages of data can be ignored during garbage collection. The SSD cannot tell what files have been deleted until the OS uses the same sectors to store new files, but by that time the SSD has already wasted cycles by garbage collecting data that was invalid, but known to the SS.

[…]

TRIM will benefit all SSDs in some way. Some will benefit more than others depending upon their situation.

[…]

Garbage collection without TRIM will always be moving all invalid data during the GC process acting like the SSD is operating at full capacity. Only the TRIM command can identify the invalid data and improve performance

[…]

GC is not a replacement for TRIM regardless of foreground or background type. TRIM can only make the GC process more efficient and lower the write amplification

- - - Auto-Merged Double Post - - -


- - - Auto-Merged Double Post - - -

Here’s the skinny: TRIM is never a requirement, but it always helps, and you’re always better off with it than without it. In this (relatively) short article, I'm going to explain why.

Now, this isn’t going to be an exhaustive look into how SSDs work. We have you covered if that’s what you’re looking for, though—just hit up this 10,000 word monster I wrote a couple of years ago if you want every single detail of what goes on inside those magical little devices. Today we're going to go through why you still get benefit from TRIM even on a modern SSD with super-efficient garbage collection. It comes down to the fact that even though SSD controllers are brilliant and getting better with every iteration, they still have no way of knowing what the operating system is doing.

Blocks and files
Any modern storage device, be it an SSD or a hard disk drive or a flash-based USB stick, has two different "layers" of internal organization. The first is the hardware layer, or the "block" layer. Blocks are usually defined during manufacture and are hard divisions—a hard disk drive’s 512-byte or 4Kbyte sector size is a block-level construction, set by the factory and programmed into the drive’s controller. On an SSD, a block is a physical construct made up of some number of pages, which are in turn made up of some number of individual NAND cells. For SSDs, a page is usually 8KB and is the smallest structure an SSD can read and write.

But Windows, OS X, and Linux can’t see into the physical structures of their disks—this has been true for decades, ever since logical block addressing became a thing. Abstracting away the physical characteristics of the disk lets the disk’s controller use lots of tricks for more efficiently organizing data behind the scenes, and this is especially true of SSDs, which have a bag of tricks that would make Penn and Teller jealous.

Instead, operating systems have their own organizational schemes—file systems. Most computer-savvy people know at least a bit about the file system they’re using—for modern Windows, it’s NTFS; OS X uses HFS+; and most Linux distros these days use ext4. Modern file systems impose their own sets of order on top of the disk’s block structures—each file system will have its own cluster size and its own way of laying out files and storing its tables and journals.

There is a barrier between the world of files and the world of blocks. The operating system doesn’t have any say in how an SSD’s controller does its job or which blocks the SSD controller uses to write data—the OS knows all about its file system, but nothing about the blocks underneath. Similarly, the SSD’s controller knows everything about what blocks are in use and what blocks are free, but it has no way of knowing which blocks correspond to which files.

And when the operating system deletes files, things get complicated.

Garbage collection—it might not mean what you think
SSDs can read and write at the page level—which is usually 8KB for modern drives—but they have a peculiar failing in that they cannot erase at the page level. SSDs can only erase entire blocks, which are usually made up of hundreds of pages. The reason for this (explained in much greater detail in our big SSD guide) is that erasing a page’s contents requires zapping that page with a not-insignificant amount of voltage, and the NAND-style layout of all modern SSDs makes it prohibitively difficult to isolate that voltage to only the pages that need erasing.

SSDs can only be erased one whole block at a time.
This is one reason why earlier SSDs seemed to slow down as they aged. It’s fast to write and fast to read, but when the disk’s initial store of free pages ran out, the disk had to erase old blocks before writing new ones.

To prevent this undesirable situation from happening, modern SSDs run complex routines called garbage collection in order to always keep as large a reserve as possible of pristine empty blocks ready for writing. Garbage collection involves having the controller search through its inventory of written pages for pages that have been marked as "stale"—that is, they were written to and then the data they contained needed to be modified by the OS; because changing the page’s state is impossible without first erasing it, the changes are always written to new pages and the old pages marked stale. Garbage collection looks for blocks that contain a mix of good and stale pages and then duplicates all the good pages into new blocks and leaves behind only stale pages in the old block. Then it erases the old block and marks it ready for use.

"Garbage collection" is therefore somewhat of a misnomer because it’s actually good pages that are "collected," with the garbage being left behind.

The complicated part comes not when a file is altered by the operating system but rather when a file is deleted. File systems don’t actually "delete" anything—when a file is deleted, the OS simply marks it in some specific way as being overwritable by new data. How this is done varies a bit by file system—NTFS, for example, updates its Master File Table to show the file system clusters occupied by the deleted file as being free for use.

But remember how the OS can’t see the SSD’s blocks and the SSD can’t see the OS’s file system? A file system deletion operation looks to the SSD like just another series of writes. On a hard disk drive, where individual sectors can be easily reused, this isn’t a problem—there’s usually a fixed correlation between the file system’s clusters and the disk’s sectors (this is, after all, what LBA was invented to facilitate). But on an SSD, where there’s no fixed correlation and where in-use pages have to be tracked and picked up by garbage collection, it can be a big deal. Pages containing deleted files look like valid pages, and they keep getting collected along with actually good pages.

Enter TRIM
This is the situation TRIM was introduced to remedy. TRIM (which is properly capitalized but is not an acronym) is an ATA command that the operating system can cause to be sent when it deletes a file. The TRIM command provides that bridge from the file level to the block level, giving the operating system a way to tell the SSD that it’s deleting files and to mark those files’ pages as stale.

With TRIM, an SSD is no longer forced to save pages belonging to deleted files. TRIM doesn’t obviate the need for garbage collection—it works with garbage collection to more properly mark pages as stale. And you don’t need TRIM for garbage collection to work—but TRIM makes an SSD’s garbage collection more efficient.


Without TRIM, garbage collection doesn't know about deleted files and continues to move pages containing deleted data along with good pages, increasing write amplification. TRIM tells the controller that it can stop collecting pages with deleted data so that they get left behind and erased with the rest of the block.
TRIM isn’t magical, and you don’t have to have it. Modern SSDs with garbage collection will work fine without it (and most SSD OEMs have utilities available to "refresh" SSDs that are being used in non-TRIM environments). Indeed, if you’re adding a third-party SSD to a Mac, you won’t be able to enable TRIM support without loading some old and insecure kernel extensions—don’t do this, because although TRIM makes a difference in reducing write amplification and extending the life and performance of an SSD, it’s not worth the barn-sized security holes you have to open up in order to get it.

So, always use TRIM if you can. It will make your SSD’s garbage collection work a lot better. But if you find yourself in a situation where TRIM isn’t available, don’t panic—it’s nice to have, but it isn’t a requirement.

View article comments

- - - Auto-Merged Double Post - - -

 

OC-NightHawk

Member
Joined
May 13, 2003
I am the Sr. Director of Marketing and Performance Analysis for the division at LSI (soon to be owned by Avago) that makes the SandForce Flash controllers. I have presented many times at the Flash Memory Summit in Santa Clara, California, on the topic of SSD operations for many years. I also have a blog on www.blog.lsi.com covering all aspects of Solid State Drives.

[…]

For the short explanation flash memory is organized in groups of pages where data can be written. Once a page is written, it cannot be rewritten until it is erased. But a page can only be erased within a group of typically 128 pages called a block. The complexity of writing data really starts to escalate in the case of random writes replacing previously written data. Random writes put the new data in previously erased pages elsewhere, peppering a block of valid data with “patches of invalid data.” In order to write new data to these patches, the whole block – all 128 pages – must be erased. But first all surrounding pages with valid data must be read and then rewritten to blank pages. The newly erased block of blank pages is then ready to save new data.

[…]

All NAND Flash-based SSDs use GC. Some use foreground GC and some use background or idle-time GC. The difference between them is covered in my blog http://blog.lsi.com/dont-let-ssds-throw-away-your-gold/ . In simple terms background garbage collection will increase write amplification (WA) and wear out the SSD sooner. Foreground GC is harder to achieve and I believe only the SandForce controller is able to do it today

[…]

TRIM is beneficial to all SSDs regardless of what kind of garbage collection is used. I talk about how TRIM came into existence and why it is necessary in my blog http://blog.lsi.com/did-you-know-hdds-do-not-have-a-del… . The TRIM command is sent by the OS to the SSD to identify what pages of data can be ignored during garbage collection. The SSD cannot tell what files have been deleted until the OS uses the same sectors to store new files, but by that time the SSD has already wasted cycles by garbage collecting data that was invalid, but known to the SS.

[…]

TRIM will benefit all SSDs in some way. Some will benefit more than others depending upon their situation.

[…]

Garbage collection without TRIM will always be moving all invalid data during the GC process acting like the SSD is operating at full capacity. Only the TRIM command can identify the invalid data and improve performance

[…]

GC is not a replacement for TRIM regardless of foreground or background type. TRIM can only make the GC process more efficient and lower the write amplification

- - - Auto-Merged Double Post - - -

https://cindori.org/trim-vs-garbage-collection/

- - - Auto-Merged Double Post - - -

Here’s the skinny: TRIM is never a requirement, but it always helps, and you’re always better off with it than without it. In this (relatively) short article, I'm going to explain why.

Now, this isn’t going to be an exhaustive look into how SSDs work. We have you covered if that’s what you’re looking for, though—just hit up this 10,000 word monster I wrote a couple of years ago if you want every single detail of what goes on inside those magical little devices. Today we're going to go through why you still get benefit from TRIM even on a modern SSD with super-efficient garbage collection. It comes down to the fact that even though SSD controllers are brilliant and getting better with every iteration, they still have no way of knowing what the operating system is doing.

Blocks and files
Any modern storage device, be it an SSD or a hard disk drive or a flash-based USB stick, has two different "layers" of internal organization. The first is the hardware layer, or the "block" layer. Blocks are usually defined during manufacture and are hard divisions—a hard disk drive’s 512-byte or 4Kbyte sector size is a block-level construction, set by the factory and programmed into the drive’s controller. On an SSD, a block is a physical construct made up of some number of pages, which are in turn made up of some number of individual NAND cells. For SSDs, a page is usually 8KB and is the smallest structure an SSD can read and write.

But Windows, OS X, and Linux can’t see into the physical structures of their disks—this has been true for decades, ever since logical block addressing became a thing. Abstracting away the physical characteristics of the disk lets the disk’s controller use lots of tricks for more efficiently organizing data behind the scenes, and this is especially true of SSDs, which have a bag of tricks that would make Penn and Teller jealous.

Instead, operating systems have their own organizational schemes—file systems. Most computer-savvy people know at least a bit about the file system they’re using—for modern Windows, it’s NTFS; OS X uses HFS+; and most Linux distros these days use ext4. Modern file systems impose their own sets of order on top of the disk’s block structures—each file system will have its own cluster size and its own way of laying out files and storing its tables and journals.

There is a barrier between the world of files and the world of blocks. The operating system doesn’t have any say in how an SSD’s controller does its job or which blocks the SSD controller uses to write data—the OS knows all about its file system, but nothing about the blocks underneath. Similarly, the SSD’s controller knows everything about what blocks are in use and what blocks are free, but it has no way of knowing which blocks correspond to which files.

And when the operating system deletes files, things get complicated.

Garbage collection—it might not mean what you think
SSDs can read and write at the page level—which is usually 8KB for modern drives—but they have a peculiar failing in that they cannot erase at the page level. SSDs can only erase entire blocks, which are usually made up of hundreds of pages. The reason for this (explained in much greater detail in our big SSD guide) is that erasing a page’s contents requires zapping that page with a not-insignificant amount of voltage, and the NAND-style layout of all modern SSDs makes it prohibitively difficult to isolate that voltage to only the pages that need erasing.

SSDs can only be erased one whole block at a time.
This is one reason why earlier SSDs seemed to slow down as they aged. It’s fast to write and fast to read, but when the disk’s initial store of free pages ran out, the disk had to erase old blocks before writing new ones.

To prevent this undesirable situation from happening, modern SSDs run complex routines called garbage collection in order to always keep as large a reserve as possible of pristine empty blocks ready for writing. Garbage collection involves having the controller search through its inventory of written pages for pages that have been marked as "stale"—that is, they were written to and then the data they contained needed to be modified by the OS; because changing the page’s state is impossible without first erasing it, the changes are always written to new pages and the old pages marked stale. Garbage collection looks for blocks that contain a mix of good and stale pages and then duplicates all the good pages into new blocks and leaves behind only stale pages in the old block. Then it erases the old block and marks it ready for use.

"Garbage collection" is therefore somewhat of a misnomer because it’s actually good pages that are "collected," with the garbage being left behind.

The complicated part comes not when a file is altered by the operating system but rather when a file is deleted. File systems don’t actually "delete" anything—when a file is deleted, the OS simply marks it in some specific way as being overwritable by new data. How this is done varies a bit by file system—NTFS, for example, updates its Master File Table to show the file system clusters occupied by the deleted file as being free for use.

But remember how the OS can’t see the SSD’s blocks and the SSD can’t see the OS’s file system? A file system deletion operation looks to the SSD like just another series of writes. On a hard disk drive, where individual sectors can be easily reused, this isn’t a problem—there’s usually a fixed correlation between the file system’s clusters and the disk’s sectors (this is, after all, what LBA was invented to facilitate). But on an SSD, where there’s no fixed correlation and where in-use pages have to be tracked and picked up by garbage collection, it can be a big deal. Pages containing deleted files look like valid pages, and they keep getting collected along with actually good pages.

Enter TRIM
This is the situation TRIM was introduced to remedy. TRIM (which is properly capitalized but is not an acronym) is an ATA command that the operating system can cause to be sent when it deletes a file. The TRIM command provides that bridge from the file level to the block level, giving the operating system a way to tell the SSD that it’s deleting files and to mark those files’ pages as stale.

With TRIM, an SSD is no longer forced to save pages belonging to deleted files. TRIM doesn’t obviate the need for garbage collection—it works with garbage collection to more properly mark pages as stale. And you don’t need TRIM for garbage collection to work—but TRIM makes an SSD’s garbage collection more efficient.


Without TRIM, garbage collection doesn't know about deleted files and continues to move pages containing deleted data along with good pages, increasing write amplification. TRIM tells the controller that it can stop collecting pages with deleted data so that they get left behind and erased with the rest of the block.
TRIM isn’t magical, and you don’t have to have it. Modern SSDs with garbage collection will work fine without it (and most SSD OEMs have utilities available to "refresh" SSDs that are being used in non-TRIM environments). Indeed, if you’re adding a third-party SSD to a Mac, you won’t be able to enable TRIM support without loading some old and insecure kernel extensions—don’t do this, because although TRIM makes a difference in reducing write amplification and extending the life and performance of an SSD, it’s not worth the barn-sized security holes you have to open up in order to get it.

So, always use TRIM if you can. It will make your SSD’s garbage collection work a lot better. But if you find yourself in a situation where TRIM isn’t available, don’t panic—it’s nice to have, but it isn’t a requirement.

View article comments

- - - Auto-Merged Double Post - - -

https://arstechnica.com/gadgets/201...garbage-collection-so-i-dont-need-trim-right/

Fascinating but maybe we should move this to it’s own topic so it can get the attention it deserves without detracting from the graphics card debate. :D
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
Not sure which is more hilarious... the article copy and paste into a forum post, or that the entire thing was then quoted... :rofl:.


Thanks, taco, for the lesson on GC and TRIM. Plenty of threads here discussing their merits already. It hasn't been an issue in several years. ;)