PS3 Project RSX Boost: Overclock your Retail PS3 RSX Speeds (ps3 cfw only)

Tanzu15 said:
I'm dying to know too. I have a 40nm super slim catching dust.
Click to expand...
iTzSweetTooth said:
thats amazing super slim can do OC now. Can you please tell us how high can the OC reach.

What exactly is BadHTAB ?! Is it a hardware mode !
Click to expand...
Details can be found here, with other links there pointing to the code and other projects. You basically need to solder 2 0.1mm magnet wires (30cm stable for me, ~40cm is original configuration from @aomsin2526 ) Solder gp15 and gp16 from pico to rq7 and rq8 (or use alt rq7, but I used original resistor points to solder to). Install the uf2 file onto pico for glitching, install the BadHTAB pkg on ps3, and that's it. Plug a USB into pico to ps3, and run the eboot from xmb.You can see the pics of my 2 consoles in this video from @MrMario2011 https://www.psx-place.com/threads/m...sor-exploit-for-ps3-xbox-360-games-get.47366/

The newest open beta for hen has support for testing clocks, although a newer build is being tested now with fix for mem clock crashing previously. Also added new functions to poke core and mem separately, with new xml items for each.

I tested on the emmc (28nm) but have not tested nor yet. The emmc one ran OK at 700/700 and lower values like 100, but the 1000/1000 crashed (only tested a couple times though)
 
Last edited:
esc0rtd3w said:
Details can be found here, with other links there pointing to the code and other projects. You basically need to solder 2 0.1mm magnet wires (30cm stable for me, ~40cm is original configuration from @aomsin2526 ) Solder gp15 and gp16 from pico to rq7 and rq8 (or use alt rq8, but I used original resistor points to solder to). Install the uf2 file onto pico for glitching, install the BadHTAB pkg on ps3, and that's it. Plug a USB into pico to ps3, and run the eboot from xmb.You can see the pics of my 2 consoles in this video from @MrMario2011 https://www.psx-place.com/threads/m...sor-exploit-for-ps3-xbox-360-games-get.47366/

The newest open beta for hen has support for testing clocks, although a newer build is being tested now with fix for mem clock crashing previously. Also added new functions to poke core and mem separately, with new xml items for each.

I tested on the emmc (28nm) but have not tested nor yet. The emmc one ran OK at 700/700 and lower values like 100, but the 1000/1000 crashed (only tested a couple times though)
Click to expand...
If 1000/1000 crashes then I guess there wasn't much of a lift going from 40 to 28. But man if this requires soldering, it's just meh.
 
Tanzu15 said:
If 1000/1000 crashes then I guess there wasn't much of a lift going from 40 to 28. But man if this requires soldering, it's just meh.
Click to expand...

Meh? This is not for you then, the goal was to know how far 28nm can go. It is all about experiment and numbers. Not daily driving it.

While 5-10% may sound scary, in reality it isn't that bad at all. Soldering is easier than RGH or switch modchip. It just 2 wires. Hard part is getting right setup to get it to boot first. But once you know how then it's easy and you can repeat it with other consoles.

5-10% is success rate PER RUN. But each run only takes 2min max. This mean you can do lot of retry rapidly. Unlike BadUpdate where each run can take hours.

This means in practice BadHTAB success rate is better than BadUpdate, you should be able to get successful run within 5-20min avg depends on your setup quality and luck. Just lots of rebooting.
 
aomsin2526 said:
Meh? This is not for you then, the goal was to know how far 28nm can go. It is all about experiment and numbers. Not daily driving it.

While 5-10% may sound scary, in reality it isn't that bad at all. Soldering is easier than RGH or switch modchip. It just 2 wires. Hard part is getting right setup to get it to boot first. But once you know how then it's easy and you can repeat it with other consoles.

5-10% is success rate PER RUN. But each run only takes 2min max. This mean you can do lot of retry rapidly. Unlike BadUpdate where each run can take hours.

This means in practice BadHTAB success rate is better than BadUpdate, you should be able to get successful run within 5-20min avg depends on your setup quality and luck. Just lots of rebooting.
Click to expand...
Hi master, sorry out of topic, is there any possibility, using this method to install CFW on console which is not CFW compatible?
 
About the new overclocking method with the xai and webmanMOD,

Does this eliminate most of the bricking possibilities when messing up with the values?

Since it reboots back with stock values and is temporary and not permanent like a update file

Sent from my SM-A556E using Tapatalk
 
RIP-Felix said:
My buddy Johnny just overclocked a Golden Goose Batch 40nm (CXD5300GGB) in his watwrcooled frankie to 1200/1050 with overvolt. It wasnt fully stable and he had to back off to 1000/1000, but that is a truely stable 1 GHz PS3. Also he could do that without overvolting.
Click to expand...
One of my ps3 slims can do 950/1000 without any voltage at all fully stable. I have a silicon winner. Pretty sure it can easily do 1000/1000 zero problems with a bit more voltage.
 
esc0rtd3w said:
Details can be found here, with other links there pointing to the code and other projects. You basically need to solder 2 0.1mm magnet wires (30cm stable for me, ~40cm is original configuration from @aomsin2526 ) Solder gp15 and gp16 from pico to rq7 and rq8 (or use alt rq7, but I used original resistor points to solder to). Install the uf2 file onto pico for glitching, install the BadHTAB pkg on ps3, and that's it. Plug a USB into pico to ps3, and run the eboot from xmb.You can see the pics of my 2 consoles in this video from @MrMario2011 https://www.psx-place.com/threads/m...sor-exploit-for-ps3-xbox-360-games-get.47366/

The newest open beta for hen has support for testing clocks, although a newer build is being tested now with fix for mem clock crashing previously. Also added new functions to poke core and mem separately, with new xml items for each.

I tested on the emmc (28nm) but have not tested nor yet. The emmc one ran OK at 700/700 and lower values like 100, but the 1000/1000 crashed (only tested a couple times though)
Click to expand...

That's great - I don't see any real detailed pics so this probably is only for super experienced PS3 modders, but I have a superslim I'd love to try at some point (bought new in 2013, don't have serial on hand). Once further detailed instructions come out I totally can try (experience with OGXbox and RGH 360). I wonder how high the 28nm RSX can go...

Also did it turn out that CXD5300GGB is the best GPU? I think that was the theory a year ago. Mine can also approach 1000 on both.
 
Last edited:
The 28nm RSX uses GDDR5 memory so i guess the code for setting the speed there is going to be different. Setting the GDDR5 to 1000 would be equivalent to double the clock vs GDDR3.
 
Last edited:
RIP-Felix said:
Update. 28nm has been overclocked! @ngx just achieved 800 core. crashed at 850. Hasn't found upper limit of VRAM yet. So the trend of TSMC made dies being crap appears to have followed the RSX into the 28nm as well. We need to find a Toshiba die and try it.
Click to expand...

Wow, that's truly unfortunate. I really thought that would be our path to 1000mhz+. Be interested in seeing the thermal efficiency, the way the heat pipe attaches to the 28nm die kinda sucks.
 
RIP-Felix said:
Update. 28nm has been overclocked! @ngx just achieved 800 core. crashed at 850. Hasn't found upper limit of VRAM yet. So the trend of TSMC made dies being crap appears to have followed the RSX into the 28nm as well. We need to find a Toshiba die and try it.
Click to expand...
Sounds like we kinda found the design limit of the chip, regarded of manufacturing.
 
WailkyS5M5 said:
About the new overclocking method with the xai and webmanMOD,

Does this eliminate most of the bricking possibilities when messing up with the values?

Since it reboots back with stock values and is temporary and not permanent like a update file

Sent from my SM-A556E using Tapatalk
Click to expand...

webMAN MOD reduces the possibilities of bricking due the frequencies are set temporarily in memory.
However it does not *eliminate* it if the new high frequencies caused a hardware damage while they were active.

I recommend to first test the allowed GPU frequencies using a web command like /gpuclock.ps3?650|750

Once the values are confirmed working, then use the values in /setup.ps3

⚠️If an invalid value is used in /setup.ps3, the console will boot with the invalid settings and semi-brick again.

To solve this issue, press L2 or R2 while the plugin is starting to reset the frequencies to the default values.

Or disable the plugin by entering to Recovery Menu or using an empty boot_plugins.txt in the root of the USB drive on PS3HEN. After boot, it will be required to delete /dev_hdd0/tmp/wm_config.bin to restore the default values and let the plugin load again.
 
aldostools said:
webMAN MOD reduces the possibilities of bricking due the frequencies are set temporarily in memory.
However it does not *eliminate* it if the new high frequencies caused a hardware damage while they were active.

I recommend to first test the allowed GPU frequencies using a web command like /gpuclock.ps3?650|750

Once the values are confirmed working, then use the values in /setup.ps3

[emoji3544]If an invalid value is used in /setup.ps3, the console will boot with the invalid settings and semi-brick again.

To solve this issue, press L2 or R2 while the plugin is starting to reset the frequencies to the default values.

Or disable the plugin by entering to Recovery Menu or using an empty boot_plugins.txt in the root of the USB drive on PS3HEN. After boot, it will be required to delete /dev_hdd0/tmp/wm_config.bin to restore the default values and let the plugin load again.
Click to expand...
Thanks for the explanation chef
 
You must log in or register to reply here.

Similar threads

Back
Top