I'm trying to improve Fmax in our design. See the archived quartus II proejct file in the attachment. The source files are also attached. In my currrent design, after compilation, Fmax can reach only 141Mhz. But my target is to reach 300mhz. Do you know how to optimize it to achieve 300mhz? The device I'm using is Arria V 5AGXBA3D4F31C4.

The usual way to improve timing is reduce the amount of logic between registers. Use tier quest to find the worst paths and reduce the logic. Then improve this paths and keep on working on the next worry paths. With afmax if 140 I suspect you have probably got large accounts of logic between the regs. For 300mhz you'll need a max of 1-2 puts between regs and extra registers around rams and dsps as these can often be bottlenecks. Summary.. Add more pipelining

Thank you! Tricky. --- Quote Start --- The usual way to improve timing is reduce the amount of logic between registers. Use tier quest to find the worst paths and reduce the logic. Then improve this paths and keep on working on the next worry paths. --- Quote End --- I'm new to time quest. Could you help to be more specific? Or some documents you can recommend?

Altera have a time quest tutorial https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0cciqfjaa&url=https%3a%2f%2fwww.altera.com%2fen_us%2fpdfs%2fliterature%2fug%2fug_tq_tutorial.pdf&ei=lpvivf_nkcfcaoq5gygh&usg=afqjcnh-52x-lt7rit91yaorkawzgdt93w&sig2=bgivq3tfcghbrwevb2vhwq&bvm=bv.92291466,d.d2s

--- Quote Start --- For 300mhz you'll need a max of 1-2 puts between regs and extra registers around rams and dsps as these can often be bottlenecks. Summary.. Add more pipelining --- Quote End --- Hello Tricky, not that clear about this statement. Could you help to explain more? In my case, do you have some ideas to optimize my code?

This is only a start: 1) You have to add an SDC file with create_clock like this create_clock -name Clk -period 3.333 [get_ports {Clk}] so that you inform Quartus that you want 300 MHz. You will get fmax 144 MHz with Cyclone IV (I have not Quartus subscription so changed target device) 2) In Analysis & Synthesis Settings: - set Speed into Optimization Technique - enable Perform WYSIWYG primitive resynthesis 3) In Physical Synthesis Optimizations set effort to Extra and Check all options under Optimize for performance With also 2) and 3) you will get fmax 180 MHz 4) Add a PLL to compensate clock (so in Normal Mode) Wrap all with a file like this (or simply add PLL in your TOP) LIBRARY ieee; USE ieee.std_logic_1164.all; LIBRARY work; ENTITY IPG1_FIP_worker IS PORT ( Reset : IN STD_LOGIC; In_Start_In : IN STD_LOGIC; In_DE_In : IN STD_LOGIC; inclk0 : IN STD_LOGIC; In_Data_In : IN STD_LOGIC_VECTOR(35 DOWNTO 0); Setting_B_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_Disables_In : IN STD_LOGIC_VECTOR(2 DOWNTO 0); Setting_Field_In : IN STD_LOGIC_VECTOR(3 DOWNTO 0); Setting_G_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_Line_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_Pattern_In : IN STD_LOGIC_VECTOR(3 DOWNTO 0); Setting_Pixel_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_R_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_TotLine_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_TotPixel_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Out_Start_Out : OUT STD_LOGIC; Out_DE_Out : OUT STD_LOGIC; Out_Data_Out : OUT STD_LOGIC_VECTOR(35 DOWNTO 0) ); END IPG1_FIP_worker; ARCHITECTURE bdf_type OF IPG1_FIP_worker IS COMPONENT top GENERIC (BurstLength : INTEGER; Intended_compiler : STRING; Intended_device_family : STRING; IPG1_Max_combine_nr : INTEGER ); PORT(Clk : IN STD_LOGIC; Reset : IN STD_LOGIC; In_Start_In : IN STD_LOGIC; In_DE_In : IN STD_LOGIC; In_Data_In : IN STD_LOGIC_VECTOR(35 DOWNTO 0); Setting_B_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_Disables_In : IN STD_LOGIC_VECTOR(2 DOWNTO 0); Setting_Field_In : IN STD_LOGIC_VECTOR(3 DOWNTO 0); Setting_G_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_Line_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_Pattern_In : IN STD_LOGIC_VECTOR(3 DOWNTO 0); Setting_Pixel_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_R_In : IN STD_LOGIC_VECTOR(11 DOWNTO 0); Setting_TotLine_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Setting_TotPixel_In : IN STD_LOGIC_VECTOR(15 DOWNTO 0); Out_Start_Out : OUT STD_LOGIC; Out_DE_Out : OUT STD_LOGIC; Out_Data_Out : OUT STD_LOGIC_VECTOR(35 DOWNTO 0) ); END COMPONENT; COMPONENT pll PORT(inclk0 : IN STD_LOGIC; c0 : OUT STD_LOGIC ); END COMPONENT; SIGNAL SYNTHESIZED_WIRE_0 : STD_LOGIC; BEGIN b2v_inst : top GENERIC MAP(BurstLength => 64, Intended_compiler => "Default", Intended_device_family => "Default", IPG1_Max_combine_nr => 8 ) PORT MAP(Clk => SYNTHESIZED_WIRE_0, Reset => Reset, In_Start_In => In_Start_In, In_DE_In => In_DE_In, In_Data_In => In_Data_In, Setting_B_In => Setting_B_In, Setting_Disables_In => Setting_Disables_In, Setting_Field_In => Setting_Field_In, Setting_G_In => Setting_G_In, Setting_Line_In => Setting_Line_In, Setting_Pattern_In => Setting_Pattern_In, Setting_Pixel_In => Setting_Pixel_In, Setting_R_In => Setting_R_In, Setting_TotLine_In => Setting_TotLine_In, Setting_TotPixel_In => Setting_TotPixel_In, Out_Start_Out => Out_Start_Out, Out_DE_Out => Out_DE_Out, Out_Data_Out => Out_Data_Out); b2v_inst1 : pll PORT MAP(inclk0 => inclk0, c0 => SYNTHESIZED_WIRE_0); END bdf_type; Note: You should use PLL compatible with your device, maybe ALTPLL is not compatible with Arria V5. You should also add PLL with Megawizard Plugin Manger to your project. I recommend you to use 50 MHz oscillator onboard and multiply it with PLL to 300 MHz. So SDC became: derive_clock_uncertainty derive_pll_clocks create_clock -name inclk0 -period 20.0 Now fmax (always on Cyclone IV but on Arria V will be better) will be 183 MHz. Slightly better but helps with timings of a lot of paths. 5) Run Timequest and start Report Timing Closure Recommendations, you will see that to improve you have to add pipeline (= Flip-Flop) between some paths.. Hottest path is from top:b2v_inst|testpattern_Setting_TotPixel[5] to top:b2v_inst|IPG1_FIP_worker_testpattern:i_testpattern|HelperGrayBar[1]_OTERM23, you have Adder and LessThan operator in same clock cycle, check in your code what is wrong and try to split it into 2 clock cycles and repeat 5)

Fmax Improvment | Altera Community

22 Replies

Altera_Forum
Honored Contributor
11 years ago
Thank you. flz47655.
One more question: What's the difference of Fmax and Restricted Fmax?
I used timequest to report the Fmax. And find there is a big differnece between Fmax and Restricted Fmax. See attached picture.
https://www.alteraforum.com/forum/attachment.php?attachmentid=10599
Altera_Forum
Honored Contributor
11 years ago
Physical pins have limits, you should go with differential signals for higher frequencies
Altera_Forum
Honored Contributor
11 years ago
--- Quote Start ---
What this code shows is how verbose it can become when you start using innapropriate types.

Why is setting_pixel not an unsigned already?
instead of shift_right function, why not do /2^n? so a 1 bit shift is /2, 2 = /2 etc.

so instead of the massive long lines you already have, you could just use:

elsif Setting_Pixel < ( Setting_TotPixel/2 + Setting_TotPixel/8 ) then

--- Quote End ---

That's a good idea. Do you think "/" operation will lead to lower Fmax than "shift_right"?
Altera_Forum
Honored Contributor
11 years ago
--- Quote Start ---
Thank you. flz47655.
One more question: What's the difference of Fmax and Restricted Fmax?
I used timequest to report the Fmax. And find there is a big differnece between Fmax and Restricted Fmax. See attached picture.
https://www.alteraforum.com/forum/attachment.php?attachmentid=10599
--- Quote End ---

FMax is a theoretical max frequency if there are no restrictions
Restricted FMax is when there are restrictions ( timing specs provided by engineer, limitations of the part you are using etc).

--- Quote Start ---

That's a good idea. Do you think "/" operation will lead to lower Fmax than "shift_right"?

--- Quote End ---

There is no different between shift_right and /2^n. They are the same operation. But using the / operator can make the code more readible.
Be aware that dividing by some other value (that is not 2^n) will instantiate a divider, and that will have a terrrible fmax).

Altera_Forum

Honored Contributor

11 years ago

Thank you. Tricky and flz476555. I already change my code with "unsigned" to make it more readable as below.

In my code, there is a long "if" + "elsif" chain which has lower Fmax very much. Do you have any idea to optimize the code if adding pipeline is not an option?

signal Setting_TotPixel : unsigned(15 downto 0) := (others=>'0');
signal Setting_Pixel 	: unsigned(15 downto 0) := (others=>'0');
		-- Gray bar(16 gray bars)
      	HelperGrayBar <= "1111"; -- gray level 15	
      	if Setting_Pixel < (Setting_TotPixel/16) then
      		HelperGrayBar <= "0000"; -- black
      	elsif Setting_Pixel < (Setting_TotPixel/8)  then
      		HelperGrayBar <= "0001"; -- gray level 1
			elsif Setting_Pixel < ((Setting_TotPixel/8)  + (Setting_TotPixel/16)) then
      		HelperGrayBar <= "0010"; -- gray level 2
      	elsif Setting_Pixel < (Setting_TotPixel/4)  then
      		HelperGrayBar <= "0011"; -- gray level 3
			elsif Setting_Pixel < ((Setting_TotPixel/4)) + ((Setting_TotPixel/16)) then
      		HelperGrayBar <= "0100"; -- gray level 4
			elsif Setting_Pixel < ((Setting_TotPixel/4)) + ((Setting_TotPixel/8)) then
      		HelperGrayBar <= "0101"; -- gray level 5
			elsif Setting_Pixel < ((Setting_TotPixel/2) - (Setting_TotPixel/16)) then
      		HelperGrayBar <= "0110"; -- gray level 6
			elsif Setting_Pixel < (Setting_TotPixel/2) then
      		HelperGrayBar <= "0111"; -- gray level 7
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/16)) then
      		HelperGrayBar <= "1000"; -- gray level 8
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/8)) then
      		HelperGrayBar <= "1001"; -- gray level 9
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/8) + (Setting_TotPixel/16)) then
      		HelperGrayBar <= "1010"; -- gray level 10
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/4)) then
      		HelperGrayBar <= "1011"; -- gray level 11
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/4) + (Setting_TotPixel/16)) then
      		HelperGrayBar <= "1100"; -- gray level 12
			elsif Setting_Pixel < ((Setting_TotPixel/2) + (Setting_TotPixel/4) + (Setting_TotPixel/8)) then
      		HelperGrayBar <= "1101"; -- gray level 13
			elsif Setting_Pixel < (Setting_TotPixel - (Setting_TotPixel/16)) then
      		HelperGrayBar <= "1110"; -- gray level 14
      	end if;

Altera_Forum
Honored Contributor
11 years ago
You could probably setup all of the compare values on the previous clock cycle, and then setup some signal so that a case statement is viable instead of the if/elseif chain. if/elsif will create a priority mux if the cases are not mutually exclusive, which is slower than a normal mux (that you get from a case statement).
Altera_Forum
Honored Contributor
11 years ago
Hello Tricky, don't understand very well on your idea. Could you be more specific?
If setup all of the compare values on the previous cycle, is it just moving the long chain "if-elsif" to the previous cycle?

Altera_Forum

Honored Contributor

11 years ago

Hello Tricky, thanks a lot for your idea. I think I already got your point and already implemented it into my code as below. It did improve Fmax a lot(60mhz).

			HelperGrayBar <= (others=>'0'); 	
			if Setting_Pixel_un < (Setting_TotPixel_un/16) then
				HelperGrayBar(0) <= '1'; 
			end if;
			
			if Setting_Pixel_un < (Setting_TotPixel_un/8) then
				HelperGrayBar(1) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/8)  + (Setting_TotPixel_un/16)) then
				HelperGrayBar(2) <= '1'; 
			end if;
			
			if Setting_Pixel_un < (Setting_TotPixel_un/4) then
				HelperGrayBar(3) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/4)) + ((Setting_TotPixel_un/16)) then
				HelperGrayBar(4) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/4)) + ((Setting_TotPixel_un/8)) then
				HelperGrayBar(5) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) - (Setting_TotPixel_un/16)) then
				HelperGrayBar(6) <= '1'; 
			end if;
			
			if Setting_Pixel_un < (Setting_TotPixel_un/2) then
				HelperGrayBar(7) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/16)) then
				HelperGrayBar(8) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/8)) then
				HelperGrayBar(9) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/8) + (Setting_TotPixel_un/16)) then
				HelperGrayBar(10) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/4)) then
				HelperGrayBar(11) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/4) + (Setting_TotPixel_un/16)) then
				HelperGrayBar(12) <= '1'; 
			end if;
			
			if Setting_Pixel_un < ((Setting_TotPixel_un/2) + (Setting_TotPixel_un/4) + (Setting_TotPixel_un/8)) then
				HelperGrayBar(13) <= '1'; 
			end if;
			
			if Setting_Pixel_un < (Setting_TotPixel_un - (Setting_TotPixel_un/16)) then
				HelperGrayBar(14) <= '1'; 
			end if;
				if HelperGrayBar(0) = '1' then  --black
					Out_Data <= GrayColor;
				end if;
				
				if HelperGrayBar(1 downto 0) = "10" then  --gray level 1
					Out_Data <= GrayColor + x"100100100";
				end if;
				
				if HelperGrayBar(2 downto 1) = "10" then  --gray level 2
					Out_Data <= GrayColor + x"200200200";
				end if;
				
				if HelperGrayBar(3 downto 2) = "10" then  --gray level 3
					Out_Data <= GrayColor + x"300300300";
				end if;
				
				if HelperGrayBar(4 downto 3) = "10" then  --gray level 4
					Out_Data <= GrayColor + x"400400400";
				end if;
				
				if HelperGrayBar(5 downto 4) = "10" then  --gray level 5
					Out_Data <= GrayColor + x"500500500";
				end if;
				
				if HelperGrayBar(6 downto 5) = "10" then  --gray level 6
					Out_Data <= GrayColor + x"600600600";
				end if;
				
				if HelperGrayBar(7 downto 6) = "10" then  --gray level 7
					Out_Data <= GrayColor + x"700700700";
				end if;
				
				if HelperGrayBar(8 downto 7) = "10" then  --gray level 8
					Out_Data <= GrayColor + x"800800800";
				end if;
				
				if HelperGrayBar(9 downto 8) = "10" then  --gray level 9
					Out_Data <= GrayColor + x"900900900";
				end if;
				
				if HelperGrayBar(10 downto 9) = "10" then  --gray level 10
					Out_Data <= GrayColor + x"A00A00A00";
				end if;
				
				if HelperGrayBar(11 downto 10) = "10" then  --gray level 11
					Out_Data <= GrayColor + x"B00B00B00";
				end if;
				
				if HelperGrayBar(12 downto 11) = "10" then  --gray level 12
					Out_Data <= GrayColor + x"C00C00C00";
				end if;
				
				if HelperGrayBar(13 downto 12) = "10" then  --gray level 13
					Out_Data <= GrayColor + x"D00D00D00";
				end if;
				
				if HelperGrayBar(14 downto 13) = "10" then  --gray level 14
					Out_Data <= GrayColor + x"E00E00E00";
				end if;
				
				if HelperGrayBar(14) = '0' then  --gray level 15
					Out_Data <= GrayColor + x"F00F00F00";
				end if;

Altera_Forum
Honored Contributor
11 years ago
This still builds a priority encoder, as in VHDL the last assignment always takes priority.

And if you're only getting an FMax of 60MHz, there are still probably some long logic paths. You should use Timequest to find the worst paths, and then try and reduce the logic.
Altera_Forum
Honored Contributor
11 years ago
--- Quote Start ---
This still builds a priority encoder, as in VHDL the last assignment always takes priority.

And if you're only getting an FMax of 60MHz, there are still probably some long logic paths. You should use Timequest to find the worst paths, and then try and reduce the logic.
--- Quote End ---

Hello Tricky, Thanks for your help. Is it possible you can show me some code to explain your idea?

Forum Discussion

Fmax Improvment

22 Replies

Recent Discussions

Agilex 3 A3CZ100BM16AE7S – 3.3 V HVIO input always reads HIGH

HSIO Differential Clock Output VOD and VOCM

Check COD for 5CSXFC6C6U23C8N , 10CX220YF780E6G

Finding .qdz Files & Help w Installer

DECA MAX 10 ADC1/ADC2 via SMA – simultaneous or multiplexed?