📄 54dsplib_errata.html
字号:
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=windows-1252">
<META NAME="Generator" CONTENT="Microsoft Word 97">
<TITLE>C54x DSPLIB Errata Page</TITLE>
<META NAME="Template" CONTENT="C:\MSOFFICE\OFFICE\html.dot">
</HEAD>
<BODY LINK="#0000ff" VLINK="#800080">
<B><FONT SIZE=5><P ALIGN="CENTER">C54x DSPLIB Errata Page</P>
</FONT><U><FONT SIZE=4><P> </P>
<P>Bug#1: ARCT2</P><DIR>
<DIR>
</U></FONT><P>Description:</P>
</B><P>This function would not return to the correct address if a calculation was performed in quadrant 1.</P>
<B><P>Solution:</P>
</B><P>The following section of code had a "frame" instruction wrongly inserted. This led to the function returning wrongly.</P>
<P>QUADRANT1:</P>
<P>	sth	a, *AR_Z+</P>
<P>	banz	LOOP, *AR_N-</P>
<P>	frame	#FSZ			/**** BUG ***/</P>
<P>	b	DONE</P>
<P>	nop</P>
<P>In version 1.1, the "frame #FSZ" instruction is replaced with a "NOP".</P>
<P> </P></DIR>
</DIR>
<B><U><FONT SIZE=4><P>Bug#2: CBREV</P>
</B></U></FONT><P> <B>Description:</P><DIR>
<DIR>
</B><P>This function would bit reverse wrongly if out-of-place processing was performed.</P></DIR>
</DIR>
<B><P>Solution:</P><DIR>
<DIR>
</B><P>The following section of code had a "ssbx sxm" instruction wrongly inserted. This led to the function sign extending the addresses and resulting in wrong reversal.</P>
<P>_cbrev</P>
<P>	ssbx	frct</P>
<P>	ssbx	sxm</P>
<P>In version 1.1, the "ssbx sxm" instruction is removed.</P>
<P> </P></DIR>
</DIR>
<B><U><FONT SIZE=4><P>Bug#3: LOG_10</P>
</B></U></FONT><FONT SIZE=2><P> </P>
</FONT><B><P>Description:</P>
</B><P>This function would return wrong results if used for far mode operation.</P>
<P> <B>Solution:</P><DIR>
<DIR>
</B><P>The number of elements, NX, was wrongly read during far mode calls because the stack offset was omitted in the LD instruction used, this can be seen below:</P>
<P> </P>
<P>**initialization for Logarithm calculation </P>
<P> </P></DIR>
</DIR>
<P>	ST	#04000h,CST_4000</P><DIR>
<DIR>
<P>	.</P></DIR>
</DIR>
<P>		.</P>
<P>		.</P>
<P>	STL	A, x_ptr</P>
<P> 	LD	*SP(17), A			/**** BUG ***/</P>
<P> 	In version 1.1, the LD instruction above is replaced with the following instruction:</P>
<P> 	LD	*SP(17+OFFSET), A</P>
<P> </P>
<B><U><FONT SIZE=4><P>Bug#4: RFFT</P>
<P> </P>
</U></FONT><P>Description:</P>
</B><P>This function would return wrong results under numerous conditions.</P>
<P> <B>Solution:</P><DIR>
<DIR>
</B><P>The unpack routine used at the end of the cfft function calls was wrong.</P></DIR>
</DIR>
<P>In version 1.1, a new unpack routine has been written.</P>
<P> </P>
<B><U><FONT SIZE=4><P>Bug#5: RIFFT</P>
</U></FONT><P>Description:</P>
</B><P>This function would return wrong results under numerous conditions.</P>
<P> <B>Solution:</P><DIR>
<DIR>
</B><P>The unpacki routine used at the end of the cifft function calls was wrong and also called wrongly.</P>
<P>In version 1.1, a new unpacki function has been written. Unpacki is called before the cifft function calls. (See examples directory for rifft usage example) </P>
<P> </P></DIR>
</DIR>
<B><U><FONT SIZE=4><P>New Functions:</P>
</B></U><P>1. CFFT32 - </FONT>32-bit fixed point complex forward FFT implementation</P>
<FONT SIZE=4><P>2. CIFFT32 - </FONT>32-bit fixed point complex inverse FFT implementation</P></BODY>
</HTML>
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -