Reply to thread
GBAtemp.net - The Independent Video Game Community
Search
Search titles only
By:
Search titles only
By:
Reply to thread
GBAtemp.net - The Independent Video Game Community
Home
Log in
Terms & Rules
Donate
Forums
New posts
Search forums
Groups
Public Events
New
New posts
New resources
New blog entries
New profile posts
New blog entry comments
New threadmarks
Latest activity
Cheats
Cheat Codes Add and Request group
The Legend of Zelda: Tears of the Kingdom cheat codes
Pokémon Legends: Arceus cheat codes
Xenoblade Chronicles 3 cheat codes
Fire Emblem Engage cheat codes
Request a cheat...
Tutorials
Nintendo Switch tutorials
Nintendo 3DS tutorials
Nintendo Wii U tutorials
Reviews
Overview
Official reviews
User reviews
Downloads
Latest reviews
Search resources
Blogs
New entries
New comments
Blog list
Search blogs
Chat
Top chatters
Search
Search titles only
By:
Search titles only
By:
Log in
Register
New posts
Search forums
Log in
Register
Home
Forums
PC, Console & Handheld Discussions
Nintendo Wii U
Fixing system memory error 160-0103 (failing eMMC) without soldering | using redNAND with ISFShax
Menu
Install the app
Install
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Reply to thread
Message
<blockquote data-quote="V10lator" data-source="post: 10410420" data-attributes="member: 485896"><p>And compared to MLC without SCFM?</p><p></p><p>Nah, seriously, I don't think anyone did benchmarks on that. The only think I heared from that is "Installing MK8 feels so much faster" (on MLC without SCFM) but that was ofc. purely subjective and not measured from IIRC [USER=636278]@SDIO[/USER].</p><p></p><p>Maybe some bored homebrew dev could write a small benchmark? Basically just fill as much RAM as possible with random data, then write that to a file on the MLC. After writing read it back into memory. Take the times both actions (writing and reading, not filling RAM with random data) took... And finally delete the file again to not waste around 1 GB of space there.</p><p>Read/write in 1 MB chunks for best performance / to max out the hardware busses.</p><p>At best don't do filesystem I/O from the UI thread but spawn a new one on a unused core (not core 2 as IIRC that one runs CafeOS stuff).</p><p>Run on Aroma but with a basic/minimal plugin set, so no plugin will do filesystem I/O in the background. Also maybe add a cooldown, like don't do anything in the first minute after app launch to give CafeOS / IOSU / Aroma time to do their things in the background before starting the benchmark.</p><p></p><p>On the other side I don't expect to see much differences. I'm not sure about that but the eMMC and the SDIO bus should run at the same clockspeed, so both should be equally fast (or slow: No need for the fastest SD card for rednand as the bus will be the bottleneck anyway, so better get something that lasts longer than default cards instead of the fastest card on the market). On the other side if you redirect SLC and stuff, too, this might slow down applications accessing multiple storages in parallell (but are there even such apps?).</p><p></p><p>Disabling SCFM might have a bigger effect through but even there I'm unsure if the speedup would be visible in benchmarks only or if it would have real-live effects... I for myself don't feel that a console with SCFM disabled is any faster, so my subjective feeling is basically the opposite of [USER=636278]@SDIO[/USER]s.</p><p></p><p>//EDIT: Writing and reading one big file with random data would btw only be benachmarking sequential reading and writing, so maybe another benchmark which would tell us more about real world situations would be random access (read/write/delete a lot of small files containing random data, maybe even from multiple threads/cores in parallel).</p></blockquote><p></p>
[QUOTE="V10lator, post: 10410420, member: 485896"] And compared to MLC without SCFM? Nah, seriously, I don't think anyone did benchmarks on that. The only think I heared from that is "Installing MK8 feels so much faster" (on MLC without SCFM) but that was ofc. purely subjective and not measured from IIRC [USER=636278]@SDIO[/USER]. Maybe some bored homebrew dev could write a small benchmark? Basically just fill as much RAM as possible with random data, then write that to a file on the MLC. After writing read it back into memory. Take the times both actions (writing and reading, not filling RAM with random data) took... And finally delete the file again to not waste around 1 GB of space there. Read/write in 1 MB chunks for best performance / to max out the hardware busses. At best don't do filesystem I/O from the UI thread but spawn a new one on a unused core (not core 2 as IIRC that one runs CafeOS stuff). Run on Aroma but with a basic/minimal plugin set, so no plugin will do filesystem I/O in the background. Also maybe add a cooldown, like don't do anything in the first minute after app launch to give CafeOS / IOSU / Aroma time to do their things in the background before starting the benchmark. On the other side I don't expect to see much differences. I'm not sure about that but the eMMC and the SDIO bus should run at the same clockspeed, so both should be equally fast (or slow: No need for the fastest SD card for rednand as the bus will be the bottleneck anyway, so better get something that lasts longer than default cards instead of the fastest card on the market). On the other side if you redirect SLC and stuff, too, this might slow down applications accessing multiple storages in parallell (but are there even such apps?). Disabling SCFM might have a bigger effect through but even there I'm unsure if the speedup would be visible in benchmarks only or if it would have real-live effects... I for myself don't feel that a console with SCFM disabled is any faster, so my subjective feeling is basically the opposite of [USER=636278]@SDIO[/USER]s. //EDIT: Writing and reading one big file with random data would btw only be benachmarking sequential reading and writing, so maybe another benchmark which would tell us more about real world situations would be random access (read/write/delete a lot of small files containing random data, maybe even from multiple threads/cores in parallel). [/QUOTE]
Insert quotes…
Verification
Post reply
Home
Forums
PC, Console & Handheld Discussions
Nintendo Wii U
Fixing system memory error 160-0103 (failing eMMC) without soldering | using redNAND with ISFShax
General chit-chat
Help
Users
Settings
Notifications
Miscellaneous
Inverse message direction
Display editor on top
Enable maximized mode
Display images as links
Hide bot messages
Hide statuses
Hide chatter list
Show messages from ignored users
Temporarily disable chat
Receive mention alerts
Sound notifications
Normal messages
Private messages
Whisper messages
Mention messages
Bot messages
Desktop notifications
Normal messages
Private messages
Whisper messages
Mention messages
Bot messages
Options
Options
View top chatters
No one is chatting at the moment.
@
K3Nv2
:
Nah that could actually bring lawsuits
Today at 9:53 AM
@
K3Nv2
:
Tempsuits
Today at 9:54 AM
@
BigOnYa
:
PollySuits
Today at 9:54 AM
@
BigOnYa
:
Your correct, Somebody would be guilty and there would be riots, then they storm the gbatemp capitol,
Today at 9:56 AM
@
K3Nv2
:
Online or not there are still certain rights that judges would have no issue handing out a warrant over
Today at 9:59 AM
@
K3Nv2
:
Just look at Kim dotcom
Today at 10:00 AM
@
BigOnYa
:
Honestly I'm scared to, from you, but ok, lemme turn on vpn, virtual machine, private browser first
Today at 10:01 AM
@
K3Nv2
:
Remember that Alexa robot I gifted you
Today at 10:04 AM
@
K3Nv2
:
And that laptop Webcam you never tapped up
Today at 10:05 AM
@
BigOnYa
:
That robot is here somewhere, I hear it moving around at night, but I haven't seen it for months.
Today at 10:08 AM
@
BigOnYa
:
Oh that laptop I give to ancientboi, so you been watching him for months, and he's been watching you
Today at 10:10 AM
@
K3Nv2
:
Oh good more than enough material for the fbi
+2
Today at 10:11 AM
@
BigOnYa
:
Damn its 5 in morn, I gotta Go wake your mum and send her to work. Check ya later.
+1
Today at 10:19 AM
@
AncientBoi
:
lol
+1
Today at 12:35 PM
@
SylverReZ
:
https://www.youtube.com/watch?v=m2KV8MHRJlQ
+2
Today at 1:42 PM
@
SylverReZ
:
@BigOnYa
,
https://www.youtube.com/watch?v=YgtO5sebA9U
+3
Today at 2:00 PM
@
Psionic Roshambo
:
He could make so much money!!! His arm would never get tired lol
+2
Today at 2:22 PM
@
SylverReZ
:
@Psionic Roshambo
, Lol
+1
Today at 2:23 PM
S
@
salazarcosplay
:
How are yall doing
+2
Today at 2:35 PM
@
VXNlcm5hbWU
:
Just had a pure banger of a sandwich there, was very nice
+1
Today at 4:11 PM
@
BakerMan
:
what do you call a dog with no legs?
53 minutes ago
@
Psionic Roshambo
:
A hot dog?
47 minutes ago
@
BakerMan
:
doesn't matter what you call it, it won't come
+1
44 minutes ago
@
SylverReZ
:
I'm having mince beef burgers with flat cheese and tomato sauce.
+1
37 minutes ago
@
BakerMan
:
34 minutes ago
Submit
@
BakerMan
:
34 minutes ago
Chat
0