专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明申请

US20180308201A1 COMPUTE OPTIMIZATION MECHANISM 审中-公开
公开(公告)号：US20180308201A1
公开(公告)日：2018-10-25
申请号：US15494905
申请日：2017-04-24
申请人： Abhishek R. Appu , Altug Koker , Linda L. Hurd , Dukhwan Kim , Mike B. Macpherson , John C. Weast , Feng Chen , Farshad Akhbari , Narayan Srinivasa , Nadathur Rajagopalan Satish , Joydeep Ray , Ping T. Tang , Michael S. Strickland , Xiaoming Chen , Anbang Yao , Tatiana Shpeisman
发明人： Abhishek R. Appu , Altug Koker , Linda L. Hurd , Dukhwan Kim , Mike B. Macpherson , John C. Weast , Feng Chen , Farshad Akhbari , Narayan Srinivasa , Nadathur Rajagopalan Satish , Joydeep Ray , Ping T. Tang , Michael S. Strickland , Xiaoming Chen , Anbang Yao , Tatiana Shpeisman
IPC分类号： G06T1/20
CPC分类号： G06T1/20 , G06F9/3001 , G06F9/3017 , G06F9/3851 , G06F9/3887 , G06F9/3895 , G06F9/46 , G06N3/063 , G06T15/005 , G06T15/04 , G09G5/363
摘要： An apparatus to facilitate compute optimization is disclosed. The apparatus includes sorting logic to sort processing threads into thread groups based on bit depth of floating point thread operations.

2. 发明授权

US07555508B2 Methods and apparatus for performing calculations using reduced-width data 有权
标题翻译：使用缩减宽度数据进行计算的方法和装置
公开(公告)号：US07555508B2
公开(公告)日：2009-06-30
申请号：US11657300
申请日：2007-01-24
申请人： Ping T. Tang , Gopi K. Kolli
发明人： Ping T. Tang , Gopi K. Kolli
IPC分类号： G06F1/00
CPC分类号： G06F7/552 , G06F7/483 , G06F2207/5523
摘要： Methods, apparatus, and articles of manufacture for performing calculations using reduced-width data are disclosed. In particular, an example method determines reduced-width data values associated with generating and evaluating functions. Some of the reduced-width data values are stored within instructions in an instruction memory during a compile phase and retrieved from instruction memory during a runtime phase.
摘要翻译：公开了使用缩减宽度数据执行计算的方法，装置和制品。特别地，示例性方法确定与生成和评估功能相关联的缩小宽度数据值。一些缩减宽度数据值在编译阶段存储在指令存储器中的指令中，并在运行阶段期间从指令存储器中检索。

3. 发明授权

US07499962B2 Enhanced fused multiply-add operation 有权
标题翻译：增强融合乘法运算
公开(公告)号：US07499962B2
公开(公告)日：2009-03-03
申请号：US11019921
申请日：2004-12-21
申请人： Ping T. Tang , David D. Donofrio
发明人： Ping T. Tang , David D. Donofrio
IPC分类号： G06F7/483
CPC分类号： G06F7/5443 , G06F7/483
摘要： An apparatus, method, and system for performing an enhanced fused multiply-add operation is disclosed. In one embodiment, an apparatus includes an exponent unit. The exponent unit includes a first adder to generate S1, where S1 is the sum of an integer k, the exponent of a floating point value A, and the exponent of a floating point value B. The exponent unit also includes a comparator to generate E1, where E1 is the greater of S1 and the exponent of a floating point value C. The apparatus also includes a partial multiplier, a shifter, and a second adder. The partial multiplier generates the partial products of the mantissas of A and B. The shifter aligns the partial products and the mantissa of C, based on E1. The second adder adds the aligned partial products and the mantissa of C. The apparatus is able to generate not only (A*B+C), but is enhanced to also be able to generate (2k*A*B+C) and the closest integer to (2k*A*B) in two's complement or floating point format.
摘要翻译：公开了一种用于执行增强的融合乘法运算的装置，方法和系统。在一个实施例中，装置包括指数单元。指数单元包括产生S1的第一加法器，其中S1是整数k，浮点值A的指数和浮点值B的指数之和。指数单元还包括产生E1的比较器其中E1是S1中的较大者和浮点值C的指数。该装置还包括部分乘法器，移位器和第二加法器。部分乘法器产生A和B的尾数的部分乘积。移位器基于E1对齐部分乘积和C的尾数。第二个加法器将对齐的部分积和C的尾数相加。该装置能够不仅产生（A * B + C），而且能够增强也能够生成（2k * A * B + C）和最接近的整数（2k * A * B）为二进制补码或浮点格式。

4. 发明授权

US07290024B2 Methods and apparatus for performing mathematical operations using scaled integers 有权
标题翻译：使用缩放整数执行数学运算的方法和装置
公开(公告)号：US07290024B2
公开(公告)日：2007-10-30
申请号：US10740086
申请日：2003-12-18
申请人： Ping T. Tang , Gopi K. Kolli
发明人： Ping T. Tang , Gopi K. Kolli
IPC分类号： G06F7/483
CPC分类号： G06F7/483 , G06F7/38
摘要： Methods, apparatus, and articles of manufacture for performing mathematical operations using scaled integers are disclosed. In particular, an example method identifies a scaled-integer value and determines a multiplier value and a scale value based on the scaled-integer value. The multiplier value is determined by extracting information from a first portion of a bitfield based on the scaled-integer value. The scale value is determined by extracting information from a second portion of the bitfield based on the scaled-integer value. The first and second portions of the bitfield are configurable to include signed integer values. The example method then performs an arithmetic operation based on the multiplier value and the scale value.
摘要翻译：公开了使用缩放整数执行数学运算的方法，装置和制品。特别地，示例性方法识别缩放整数值，并且基于缩放整数值来确定乘数值和比例值。通过基于缩放整数值从位字段的第一部分提取信息来确定乘数值。通过基于缩放整数值从位域的第二部分提取信息来确定比例值。位域的第一和第二部分可配置为包括有符号整数值。然后，示例方法基于乘数值和比例值执行算术运算。

5. 发明授权

US07979486B2 Methods and apparatus for extracting integer remainders 有权
标题翻译：用于提取整数余数的方法和装置
公开(公告)号：US07979486B2
公开(公告)日：2011-07-12
申请号：US11946593
申请日：2007-11-28
申请人： John R. Harrison , Ping T. Tang
发明人： John R. Harrison , Ping T. Tang
IPC分类号： G06F7/535
CPC分类号： G06F7/535 , G06F7/72 , G06F2207/5356
摘要： Methods and apparatus to determine a remainder value are disclosed. A disclosed example method involves, during a compilation phase, causing a processor to multiply a dividend value by a first value to generate a second value associated with a product. The first value is associated with a scaled approximate reciprocal of a divisor value, and the scaled approximate reciprocal of the divisor value is determined using a compound exponent value. During a runtime phase, the processor is caused to multiply a third value from the second value. The third value is generated using at least a subset bitfield of the second value. During the runtime phase, the processor is caused to determine a remainder value based on the third value. The processor is caused to store the remainder value in a memory.
摘要翻译：公开了确定余数值的方法和装置。所公开的示例性方法涉及在编译阶段期间使处理器将分红值乘以第一值以生成与产品相关联的第二值。第一个值与除数值的缩放近似倒数相关联，并且使用复合指数值确定除数值的缩放近似倒数。在运行阶段期间，使处理器从第二个值乘以第三个值。使用第二值的至少子集位字段来生成第三值。在运行阶段期间，使处理器基于第三值来确定余数值。使处理器将余数值存储在存储器中。

6. 发明申请

US20090265408A1 METHODS AND APPARATUS FOR PERFORMING CALCULATIONS USING REDUCED-WIDTH DATA 有权
标题翻译：使用减少宽度数据执行计算的方法和装置
公开(公告)号：US20090265408A1
公开(公告)日：2009-10-22
申请号：US12492961
申请日：2009-06-26
申请人： Ping T. Tang , Gopi K. Kolli
发明人： Ping T. Tang , Gopi K. Kolli
IPC分类号： G06F17/11 , G06F7/38 , G06F1/02
CPC分类号： G06F7/552 , G06F7/483 , G06F2207/5523
摘要： Methods, apparatus, and articles of manufacture for performing calculations using reduced-width data are disclosed. In particular, an example method determines reduced-width data values associated with generating and evaluating functions. Some of the reduced-width data values are stored within instructions in an instruction memory during a compile phase and retrieved from instruction memory during a runtime phase.
摘要翻译：公开了使用缩减宽度数据执行计算的方法，装置和制品。特别地，示例性方法确定与生成和评估功能相关联的缩小宽度数据值。一些缩减宽度数据值在编译阶段存储在指令存储器中的指令中，并在运行阶段期间从指令存储器中检索。

7. 发明授权

US07325022B2 Methods and apparatus for determining approximating polynomials using instruction-embedded coefficients 失效
标题翻译：使用指令嵌入系数确定近似多项式的方法和装置
公开(公告)号：US07325022B2
公开(公告)日：2008-01-29
申请号：US10648500
申请日：2003-08-26
申请人： Ping T. Tang , Gopi K. Kolli , Minda Zhang
发明人： Ping T. Tang , Gopi K. Kolli , Minda Zhang
IPC分类号： G06F1/02
CPC分类号： G06F7/552 , G06F9/30014 , G06F9/3016 , G06F2207/5523
摘要： Methods and apparatus for determining approximating polynomials using instruction-embedded coefficients are disclosed. In particular, the methods and apparatus use a plurality of coefficient values stored in a plurality of instructions. The coefficient values are associated with a runtime approximating polynomial of a K-th root family function. The coefficient values and the instructions stored in an instruction memory enable the processor system to determine a K-th root family function approximation value based on the runtime approximating polynomial.
摘要翻译：公开了使用指令嵌入系数确定近似多项式的方法和装置。特别地，所述方法和装置使用存储在多个指令中的多个系数值。系数值与第K个根系函数的运行时近似多项式相关联。存储在指令存储器中的系数值和指令使得处理器系统能够基于运行时近似多项式来确定第K个根系函数近似值。

8. 发明授权

US07321916B2 Methods and apparatus for extracting integer remainders 失效
标题翻译：用于提取整数余数的方法和装置
公开(公告)号：US07321916B2
公开(公告)日：2008-01-22
申请号：US10628811
申请日：2003-07-28
申请人： John R. Harrison , Ping T. Tang
发明人： John R. Harrison , Ping T. Tang
IPC分类号： G06F7/535
CPC分类号： G06F7/535 , G06F7/72 , G06F2207/5356
摘要： Methods and apparatus for determining a remainder value are disclosed. The methods and apparatus extract a residuary subset bitfield value from a binary value that is calculated using a scaled approximate reciprocal value that is associated with a compound exponent scaling value. The residuary subset bitfield value is part of a range of contiguous bits that is associated with upper and lower boundary bit-position values that are part of the compound exponent scaling value. The methods and apparatus determine the remainder value based on the residuary subset bitfield value.
摘要翻译：公开了用于确定余数值的方法和装置。方法和装置从使用与复合指数缩放值相关联的缩放近似倒数值计算的二进制值提取剩余子集位域值。剩余子集位字段值是与作为复合指数缩放值的一部分的上边界位和下边界位值相关联的连续位的范围的一部分。方法和装置基于剩余子集位域值来确定余数值。

9. 发明授权

US07080364B2 Methods and apparatus for compiling a transcendental floating-point operation 失效
标题翻译：用于编译超验浮点运算的方法和装置
公开(公告)号：US07080364B2
公开(公告)日：2006-07-18
申请号：US10424600
申请日：2003-04-28
申请人： Ping T. Tang , Cristina S. Iordache
发明人： Ping T. Tang , Cristina S. Iordache
IPC分类号： G06F9/45
CPC分类号： G06F7/483 , G06F7/556
摘要： Methods and an apparatus for compiling a transcendental floating-point operation are disclosed. The disclosed techniques compile a transcendental floating-point operation by replacing the transcendental floating-point operation with an integer-based routine (e.g., a single routine) hand-coded to perform the transcendental floating-point operation. Each of the instructions in the integer-based routine, including the integer operations, is compiled directly into opcodes without primitive floating-point emulation calls. As a result, function nesting is reduced and more efficient algorithms are used. The disclosed system does not simply emulate basic floating-point operations using integer operations. Instead, portions of the computation are isolated where fixed-point accuracy is sufficient and thus native integer computations can be used. For example, computing (log(1+Z)−Z)/Z instead of computing log(1+Z).
摘要翻译：公开了用于编译超验浮点运算的方法和装置。所公开的技术通过用手工编码的基于整数的例程（例如，单个例程）代替先验浮点运算来编译超验浮点运算，以执行超验浮点运算。基于整数的例程中的每个指令（包括整数运算）都直接编译成没有原始浮点仿真调用的操作码。因此，功能嵌套被减少，并且使用更有效的算法。所公开的系统不仅仅使用整数运算来模拟基本的浮点运算。相反，计算的一部分被隔离，其中定点精度是足够的，因此可以使用本机整数计算。例如，计算（log（1 + Z）-Z）/ Z而不是计算log（1 + Z）。

10. 发明授权

US07031885B2 Method and apparatus for measuring accuracies of fixed-point computations 失效
标题翻译：用于测量定点计算精度的方法和装置
公开(公告)号：US07031885B2
公开(公告)日：2006-04-18
申请号：US10608244
申请日：2003-06-27
申请人： Ping T. Tang
发明人： Ping T. Tang
IPC分类号： G06F17/18 , G06F1/00
CPC分类号： G06F17/14 , G06F7/38
摘要： Method and apparatus for measuring accuracies of fixed-point computations are described herein. In one embodiment, the exemplary process includes identifying an operation involving a plurality of input values, invoking an object associated with the identified operation, performing the operation via the invoked object, and automatically generating, as a part of the operation via the object, accuracy information associated with the operation for each of the plurality of input values. Other methods and apparatuses are also described.
摘要翻译：本文描述了用于测量定点计算精度的方法和装置。在一个实施例中，示例性过程包括识别涉及多个输入值的操作，调用与所识别的操作相关联的对象，经由被调用的对象执行操作，并且通过对象自动生成作为操作的一部分的精度与多个输入值中的每一个的操作相关联的信息。还描述了其它方法和装置。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式